Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
// Copyright 2021 The Prometheus Authors
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package histogram
import (
"fmt"
"math"
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
"testing"
"github.com/stretchr/testify/require"
"github.com/prometheus/prometheus/model/value"
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
)
func TestHistogramString ( t * testing . T ) {
cases := [ ] struct {
histogram Histogram
expectedString string
} {
{
histogram : Histogram {
Schema : 0 ,
} ,
expectedString : "{count:0, sum:0}" ,
} ,
{
histogram : Histogram {
Schema : 0 ,
Count : 9 ,
Sum : - 3.1415 ,
ZeroCount : 12 ,
ZeroThreshold : 0.001 ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 5 } ,
{ Offset : 1 , Length : 1 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedString : "{count:9, sum:-3.1415, [-64,-32):1, [-16,-8):1, [-8,-4):2, [-4,-2):1, [-2,-1):3, [-1,-0.5):1, [-0.001,0.001]:12}" ,
} ,
{
histogram : Histogram {
Schema : 0 ,
Count : 19 ,
Sum : 2.7 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 0 , Length : 0 } ,
{ Offset : 0 , Length : 3 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 5 } ,
{ Offset : 1 , Length : 0 } ,
{ Offset : 0 , Length : 1 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedString : "{count:19, sum:2.7, [-64,-32):1, [-16,-8):1, [-8,-4):2, [-4,-2):1, [-2,-1):3, [-1,-0.5):1, (0.5,1]:1, (1,2]:3, (2,4]:1, (4,8]:2, (8,16]:1, (16,32]:1, (32,64]:1}" ,
} ,
}
for i , c := range cases {
t . Run ( fmt . Sprintf ( "%d" , i ) , func ( t * testing . T ) {
actualString := c . histogram . String ( )
require . Equal ( t , c . expectedString , actualString )
} )
}
}
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
func TestCumulativeBucketIterator ( t * testing . T ) {
cases := [ ] struct {
histogram Histogram
expectedBuckets [ ] Bucket [ uint64 ]
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} {
{
histogram : Histogram {
Schema : 0 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 2 } ,
{ Offset : 1 , Length : 2 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 1 , - 1 , 0 } ,
} ,
expectedBuckets : [ ] Bucket [ uint64 ] {
{ Lower : math . Inf ( - 1 ) , Upper : 1 , Count : 1 , LowerInclusive : true , UpperInclusive : true , Index : 0 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 2 , Count : 3 , LowerInclusive : true , UpperInclusive : true , Index : 1 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 4 , Count : 3 , LowerInclusive : true , UpperInclusive : true , Index : 2 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 8 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : 3 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 16 , Count : 5 , LowerInclusive : true , UpperInclusive : true , Index : 4 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} ,
} ,
{
histogram : Histogram {
Schema : 0 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 5 } ,
{ Offset : 1 , Length : 1 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedBuckets : [ ] Bucket [ uint64 ] {
{ Lower : math . Inf ( - 1 ) , Upper : 1 , Count : 1 , LowerInclusive : true , UpperInclusive : true , Index : 0 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 2 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : 1 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 4 , Count : 5 , LowerInclusive : true , UpperInclusive : true , Index : 2 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 8 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 3 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 16 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 4 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 32 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 5 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 64 , Count : 9 , LowerInclusive : true , UpperInclusive : true , Index : 6 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} ,
} ,
{
histogram : Histogram {
Schema : 0 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 7 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 0 } ,
} ,
expectedBuckets : [ ] Bucket [ uint64 ] {
{ Lower : math . Inf ( - 1 ) , Upper : 1 , Count : 1 , LowerInclusive : true , UpperInclusive : true , Index : 0 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 2 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : 1 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 4 , Count : 5 , LowerInclusive : true , UpperInclusive : true , Index : 2 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 8 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 3 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 16 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 4 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 32 , Count : 9 , LowerInclusive : true , UpperInclusive : true , Index : 5 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 64 , Count : 10 , LowerInclusive : true , UpperInclusive : true , Index : 6 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} ,
} ,
{
histogram : Histogram {
Schema : 3 ,
PositiveSpans : [ ] Span {
{ Offset : - 5 , Length : 2 } , // -5 -4
{ Offset : 2 , Length : 3 } , // -1 0 1
{ Offset : 2 , Length : 2 } , // 4 5
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 3 } ,
} ,
expectedBuckets : [ ] Bucket [ uint64 ] {
{ Lower : math . Inf ( - 1 ) , Upper : 0.6484197773255048 , Count : 1 , LowerInclusive : true , UpperInclusive : true , Index : - 5 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 0.7071067811865475 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : - 4 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 0.7711054127039704 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : - 3 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 0.8408964152537144 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : - 2 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 0.9170040432046711 , Count : 5 , LowerInclusive : true , UpperInclusive : true , Index : - 1 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 0 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1.0905077326652577 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 1 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 1.189207115002721 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 2 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1.2968395546510096 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 3 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 1.414213562373095 , Count : 9 , LowerInclusive : true , UpperInclusive : true , Index : 4 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1.5422108254079407 , Count : 13 , LowerInclusive : true , UpperInclusive : true , Index : 5 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} ,
} ,
{
histogram : Histogram {
Schema : - 2 ,
PositiveSpans : [ ] Span {
{ Offset : - 2 , Length : 4 } , // -2 -1 0 1
{ Offset : 2 , Length : 2 } , // 4 5
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedBuckets : [ ] Bucket [ uint64 ] {
{ Lower : math . Inf ( - 1 ) , Upper : 0.00390625 , Count : 1 , LowerInclusive : true , UpperInclusive : true , Index : - 2 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 0.0625 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : - 1 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1 , Count : 5 , LowerInclusive : true , UpperInclusive : true , Index : 0 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 16 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 1 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 256 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 2 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 4096 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 3 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
{ Lower : math . Inf ( - 1 ) , Upper : 65536 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 4 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1048576 , Count : 9 , LowerInclusive : true , UpperInclusive : true , Index : 5 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} ,
} ,
{
histogram : Histogram {
Schema : - 1 ,
PositiveSpans : [ ] Span {
{ Offset : - 2 , Length : 5 } , // -2 -1 0 1 2
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 } ,
} ,
expectedBuckets : [ ] Bucket [ uint64 ] {
{ Lower : math . Inf ( - 1 ) , Upper : 0.0625 , Count : 1 , LowerInclusive : true , UpperInclusive : true , Index : - 2 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 0.25 , Count : 4 , LowerInclusive : true , UpperInclusive : true , Index : - 1 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 1 , Count : 5 , LowerInclusive : true , UpperInclusive : true , Index : 0 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 4 , Count : 7 , LowerInclusive : true , UpperInclusive : true , Index : 1 } ,
{ Lower : math . Inf ( - 1 ) , Upper : 16 , Count : 8 , LowerInclusive : true , UpperInclusive : true , Index : 2 } ,
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} ,
} ,
}
for i , c := range cases {
t . Run ( fmt . Sprintf ( "%d" , i ) , func ( t * testing . T ) {
it := c . histogram . CumulativeBucketIterator ( )
actualBuckets := make ( [ ] Bucket [ uint64 ] , 0 , len ( c . expectedBuckets ) )
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
for it . Next ( ) {
actualBuckets = append ( actualBuckets , it . At ( ) )
}
require . Equal ( t , c . expectedBuckets , actualBuckets )
} )
}
}
func TestRegularBucketIterator ( t * testing . T ) {
cases := [ ] struct {
histogram Histogram
expectedPositiveBuckets [ ] Bucket [ uint64 ]
expectedNegativeBuckets [ ] Bucket [ uint64 ]
} {
{
histogram : Histogram {
Schema : 0 ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] { } ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] { } ,
} ,
{
histogram : Histogram {
Schema : 0 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 2 } ,
{ Offset : 1 , Length : 2 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 1 , - 1 , 0 } ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] {
{ Lower : 0.5 , Upper : 1 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 0 } ,
{ Lower : 1 , Upper : 2 , Count : 2 , LowerInclusive : false , UpperInclusive : true , Index : 1 } ,
{ Lower : 4 , Upper : 8 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 3 } ,
{ Lower : 8 , Upper : 16 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 4 } ,
} ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] { } ,
} ,
{
histogram : Histogram {
Schema : 0 ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 5 } ,
{ Offset : 1 , Length : 1 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] { } ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] {
{ Lower : - 1 , Upper : - 0.5 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 0 } ,
{ Lower : - 2 , Upper : - 1 , Count : 3 , LowerInclusive : true , UpperInclusive : false , Index : 1 } ,
{ Lower : - 4 , Upper : - 2 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 2 } ,
{ Lower : - 8 , Upper : - 4 , Count : 2 , LowerInclusive : true , UpperInclusive : false , Index : 3 } ,
{ Lower : - 16 , Upper : - 8 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 4 } ,
{ Lower : - 64 , Upper : - 32 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 6 } ,
} ,
} ,
{
histogram : Histogram {
Schema : 0 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 0 , Length : 0 } ,
{ Offset : 0 , Length : 3 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 5 } ,
{ Offset : 1 , Length : 0 } ,
{ Offset : 0 , Length : 1 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] {
{ Lower : 0.5 , Upper : 1 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 0 } ,
{ Lower : 1 , Upper : 2 , Count : 3 , LowerInclusive : false , UpperInclusive : true , Index : 1 } ,
{ Lower : 2 , Upper : 4 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 2 } ,
{ Lower : 4 , Upper : 8 , Count : 2 , LowerInclusive : false , UpperInclusive : true , Index : 3 } ,
{ Lower : 8 , Upper : 16 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 4 } ,
{ Lower : 16 , Upper : 32 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 5 } ,
{ Lower : 32 , Upper : 64 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 6 } ,
} ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] {
{ Lower : - 1 , Upper : - 0.5 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 0 } ,
{ Lower : - 2 , Upper : - 1 , Count : 3 , LowerInclusive : true , UpperInclusive : false , Index : 1 } ,
{ Lower : - 4 , Upper : - 2 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 2 } ,
{ Lower : - 8 , Upper : - 4 , Count : 2 , LowerInclusive : true , UpperInclusive : false , Index : 3 } ,
{ Lower : - 16 , Upper : - 8 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 4 } ,
{ Lower : - 64 , Upper : - 32 , Count : 1 , LowerInclusive : true , UpperInclusive : false , Index : 6 } ,
} ,
} ,
{
histogram : Histogram {
Schema : 3 ,
PositiveSpans : [ ] Span {
{ Offset : - 5 , Length : 2 } , // -5 -4
{ Offset : 2 , Length : 3 } , // -1 0 1
{ Offset : 2 , Length : 2 } , // 4 5
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 3 } ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] {
{ Lower : 0.5946035575013605 , Upper : 0.6484197773255048 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : - 5 } ,
{ Lower : 0.6484197773255048 , Upper : 0.7071067811865475 , Count : 3 , LowerInclusive : false , UpperInclusive : true , Index : - 4 } ,
{ Lower : 0.8408964152537144 , Upper : 0.9170040432046711 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : - 1 } ,
{ Lower : 0.9170040432046711 , Upper : 1 , Count : 2 , LowerInclusive : false , UpperInclusive : true , Index : 0 } ,
{ Lower : 1 , Upper : 1.0905077326652577 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 1 } ,
{ Lower : 1.2968395546510096 , Upper : 1.414213562373095 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 4 } ,
{ Lower : 1.414213562373095 , Upper : 1.5422108254079407 , Count : 4 , LowerInclusive : false , UpperInclusive : true , Index : 5 } ,
} ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] { } ,
} ,
{
histogram : Histogram {
Schema : - 2 ,
PositiveSpans : [ ] Span {
{ Offset : - 2 , Length : 4 } , // -2 -1 0 1
{ Offset : 2 , Length : 2 } , // 4 5
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] {
{ Lower : 0.000244140625 , Upper : 0.00390625 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : - 2 } ,
{ Lower : 0.00390625 , Upper : 0.0625 , Count : 3 , LowerInclusive : false , UpperInclusive : true , Index : - 1 } ,
{ Lower : 0.0625 , Upper : 1 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 0 } ,
{ Lower : 1 , Upper : 16 , Count : 2 , LowerInclusive : false , UpperInclusive : true , Index : 1 } ,
{ Lower : 4096 , Upper : 65536 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 4 } ,
{ Lower : 65536 , Upper : 1048576 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 5 } ,
} ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] { } ,
} ,
{
histogram : Histogram {
Schema : - 1 ,
PositiveSpans : [ ] Span {
{ Offset : - 2 , Length : 5 } , // -2 -1 0 1 2
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 } ,
} ,
expectedPositiveBuckets : [ ] Bucket [ uint64 ] {
{ Lower : 0.015625 , Upper : 0.0625 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : - 2 } ,
{ Lower : 0.0625 , Upper : 0.25 , Count : 3 , LowerInclusive : false , UpperInclusive : true , Index : - 1 } ,
{ Lower : 0.25 , Upper : 1 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 0 } ,
{ Lower : 1 , Upper : 4 , Count : 2 , LowerInclusive : false , UpperInclusive : true , Index : 1 } ,
{ Lower : 4 , Upper : 16 , Count : 1 , LowerInclusive : false , UpperInclusive : true , Index : 2 } ,
} ,
expectedNegativeBuckets : [ ] Bucket [ uint64 ] { } ,
} ,
}
for i , c := range cases {
t . Run ( fmt . Sprintf ( "%d" , i ) , func ( t * testing . T ) {
it := c . histogram . PositiveBucketIterator ( )
actualPositiveBuckets := make ( [ ] Bucket [ uint64 ] , 0 , len ( c . expectedPositiveBuckets ) )
for it . Next ( ) {
actualPositiveBuckets = append ( actualPositiveBuckets , it . At ( ) )
}
require . Equal ( t , c . expectedPositiveBuckets , actualPositiveBuckets )
it = c . histogram . NegativeBucketIterator ( )
actualNegativeBuckets := make ( [ ] Bucket [ uint64 ] , 0 , len ( c . expectedNegativeBuckets ) )
for it . Next ( ) {
actualNegativeBuckets = append ( actualNegativeBuckets , it . At ( ) )
}
require . Equal ( t , c . expectedNegativeBuckets , actualNegativeBuckets )
Style cleanup of all the changes in sparsehistogram so far
A lot of this code was hacked together, literally during a
hackathon. This commit intends not to change the code substantially,
but just make the code obey the usual style practices.
A (possibly incomplete) list of areas:
* Generally address linter warnings.
* The `pgk` directory is deprecated as per dev-summit. No new packages should
be added to it. I moved the new `pkg/histogram` package to `model`
anticipating what's proposed in #9478.
* Make the naming of the Sparse Histogram more consistent. Including
abbreviations, there were just too many names for it: SparseHistogram,
Histogram, Histo, hist, his, shs, h. The idea is to call it "Histogram" in
general. Only add "Sparse" if it is needed to avoid confusion with
conventional Histograms (which is rare because the TSDB really has no notion
of conventional Histograms). Use abbreviations only in local scope, and then
really abbreviate (not just removing three out of seven letters like in
"Histo"). This is in the spirit of
https://github.com/golang/go/wiki/CodeReviewComments#variable-names
* Several other minor name changes.
* A lot of formatting of doc comments. For one, following
https://github.com/golang/go/wiki/CodeReviewComments#comment-sentences
, but also layout question, anticipating how things will look like
when rendered by `godoc` (even where `godoc` doesn't render them
right now because they are for unexported types or not a doc comment
at all but just a normal code comment - consistency is queen!).
* Re-enabled `TestQueryLog` and `TestEndopints` (they pass now,
leaving them disabled was presumably an oversight).
* Bucket iterator for histogram.Histogram is now created with a
method.
* HistogramChunk.iterator now allows iterator recycling. (I think
@dieterbe only commented it out because he was confused by the
question in the comment.)
* HistogramAppender.Append panics now because we decided to treat
staleness marker differently.
Signed-off-by: beorn7 <beorn@grafana.com>
3 years ago
} )
}
}
func TestHistogramToFloat ( t * testing . T ) {
h := Histogram {
Schema : 3 ,
Count : 61 ,
Sum : 2.7 ,
ZeroThreshold : 0.1 ,
ZeroCount : 42 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 0 , Length : 0 } ,
{ Offset : 0 , Length : 3 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 5 } ,
{ Offset : 1 , Length : 0 } ,
{ Offset : 0 , Length : 1 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
}
fh := h . ToFloat ( )
require . Equal ( t , h . String ( ) , fh . String ( ) )
}
// TestHistogramEquals tests both Histogram and FloatHistogram.
func TestHistogramEquals ( t * testing . T ) {
h1 := Histogram {
Schema : 3 ,
Count : 61 ,
Sum : 2.7 ,
ZeroThreshold : 0.1 ,
ZeroCount : 42 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 10 , Length : 3 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 10 , Length : 3 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 , 0 } ,
}
equals := func ( h1 , h2 Histogram ) {
require . True ( t , h1 . Equals ( & h2 ) )
require . True ( t , h2 . Equals ( & h1 ) )
h1f , h2f := h1 . ToFloat ( ) , h2 . ToFloat ( )
require . True ( t , h1f . Equals ( h2f ) )
require . True ( t , h2f . Equals ( h1f ) )
}
notEquals := func ( h1 , h2 Histogram ) {
require . False ( t , h1 . Equals ( & h2 ) )
require . False ( t , h2 . Equals ( & h1 ) )
h1f , h2f := h1 . ToFloat ( ) , h2 . ToFloat ( )
require . False ( t , h1f . Equals ( h2f ) )
require . False ( t , h2f . Equals ( h1f ) )
}
h2 := h1 . Copy ( )
equals ( h1 , * h2 )
// Changed spans but same layout.
h2 . PositiveSpans = append ( h2 . PositiveSpans , Span { Offset : 5 } )
h2 . NegativeSpans = append ( h2 . NegativeSpans , Span { Offset : 2 } )
equals ( h1 , * h2 )
// Adding empty spans in between.
h2 . PositiveSpans [ 1 ] . Offset = 6
h2 . PositiveSpans = [ ] Span {
h2 . PositiveSpans [ 0 ] ,
{ Offset : 1 } ,
{ Offset : 3 } ,
h2 . PositiveSpans [ 1 ] ,
h2 . PositiveSpans [ 2 ] ,
}
h2 . NegativeSpans [ 1 ] . Offset = 5
h2 . NegativeSpans = [ ] Span {
h2 . NegativeSpans [ 0 ] ,
{ Offset : 2 } ,
{ Offset : 3 } ,
h2 . NegativeSpans [ 1 ] ,
h2 . NegativeSpans [ 2 ] ,
}
equals ( h1 , * h2 )
// All mismatches.
notEquals ( h1 , Histogram { } )
h2 . Schema = 1
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . Count ++
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . Sum ++
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . ZeroThreshold ++
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . ZeroCount ++
notEquals ( h1 , * h2 )
// Changing value of buckets.
h2 = h1 . Copy ( )
h2 . PositiveBuckets [ len ( h2 . PositiveBuckets ) - 1 ] ++
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . NegativeBuckets [ len ( h2 . NegativeBuckets ) - 1 ] ++
notEquals ( h1 , * h2 )
// Changing bucket layout.
h2 = h1 . Copy ( )
h2 . PositiveSpans [ 1 ] . Offset ++
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . NegativeSpans [ 1 ] . Offset ++
notEquals ( h1 , * h2 )
// Adding an empty bucket.
h2 = h1 . Copy ( )
h2 . PositiveSpans [ 0 ] . Offset --
h2 . PositiveSpans [ 0 ] . Length ++
h2 . PositiveBuckets = append ( [ ] int64 { 0 } , h2 . PositiveBuckets ... )
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . NegativeSpans [ 0 ] . Offset --
h2 . NegativeSpans [ 0 ] . Length ++
h2 . NegativeBuckets = append ( [ ] int64 { 0 } , h2 . NegativeBuckets ... )
notEquals ( h1 , * h2 )
// Adding new bucket.
h2 = h1 . Copy ( )
h2 . PositiveSpans = append ( h2 . PositiveSpans , Span {
Offset : 1 ,
Length : 1 ,
} )
h2 . PositiveBuckets = append ( h2 . PositiveBuckets , 1 )
notEquals ( h1 , * h2 )
h2 = h1 . Copy ( )
h2 . NegativeSpans = append ( h2 . NegativeSpans , Span {
Offset : 1 ,
Length : 1 ,
} )
h2 . NegativeBuckets = append ( h2 . NegativeBuckets , 1 )
notEquals ( h1 , * h2 )
// Sum is StaleNaN.
hStale := h1 . Copy ( )
hStale . Sum = math . Float64frombits ( value . StaleNaN )
notEquals ( h1 , * hStale )
equals ( * hStale , * hStale )
// Sum is NaN (but not a StaleNaN).
hNaN := h1 . Copy ( )
hNaN . Sum = math . NaN ( )
notEquals ( h1 , * hNaN )
equals ( * hNaN , * hNaN )
// Sum StaleNaN vs regular NaN.
notEquals ( * hStale , * hNaN )
}
histograms: Add Compact method to the normal integer Histogram
And use the new method to call to compact Histograms during
parsing. This happens for both `Histogram` and `FloatHistogram`. In
this way, if targets decide to optimize the exposition size by merging
spans with empty buckets in between, we still get a normalized
results. It will also normalize away any valid but weird
representations like empty spans, spans with offset zero, and empty
buckets at the start or end of a span.
The implementation seemed easy at first as it just turns the
`compactBuckets` helper into a generic function (which now got its own
file). However, the integer Histograms have delta buckets instead of
absolute buckets, which had to be treated specially in the generic
`compactBuckets` function. To make sure it works, I have added plenty
of explicit tests for `Histogram` in addition to the `FloatHistogram`
tests.
I have also updated the doc comment for the `Compact` method.
Based on the insights now expressed in the doc comment, compacting
with a maxEmptyBuckets > 0 is rarely useful. Therefore, this commit
also sets the value to 0 in the two cases we were using 3 so far. We
might still want to reconsider, so I don't want to remove the
maxEmptyBuckets parameter right now.
Signed-off-by: beorn7 <beorn@grafana.com>
2 years ago
func TestHistogramCompact ( t * testing . T ) {
cases := [ ] struct {
name string
in * Histogram
maxEmptyBuckets int
expected * Histogram
} {
{
"empty histogram" ,
& Histogram { } ,
0 ,
& Histogram { } ,
} ,
{
"nothing should happen" ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 1 } , { 2 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 } ,
NegativeSpans : [ ] Span { { 3 , 2 } , { 3 , 2 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , 1.234e5 , 1000 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 1 } , { 2 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 } ,
NegativeSpans : [ ] Span { { 3 , 2 } , { 3 , 2 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , 1.234e5 , 1000 } ,
} ,
} ,
{
"eliminate zero offsets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 1 } , { 0 , 3 } , { 0 , 1 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 0 , 2 } , { 2 , 1 } , { 0 , 1 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , 1.234e5 , 1000 , 3 , 4 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 5 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 4 } , { 2 , 2 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , 1.234e5 , 1000 , 3 , 4 } ,
} ,
} ,
{
"eliminate zero length" ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 2 , 0 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 0 , 0 } , { 2 , 0 } , { 1 , 4 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , 1.234e5 , 1000 , 3 , 4 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 5 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 4 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , 1.234e5 , 1000 , 3 , 4 } ,
} ,
} ,
{
"eliminate multiple zero length spans" ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 2 , 0 } , { 2 , 0 } , { 2 , 0 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 9 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
} ,
} ,
{
"cut empty buckets at start or end" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 4 } , { 5 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , 3 , 4 , - 9 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 5 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 4 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , 3 , 4 } ,
} ,
} ,
{
"cut empty buckets at start and end" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 4 } , { 5 , 6 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , 3 , - 3 , 42 , 3 , - 46 , 0 , 0 } ,
NegativeSpans : [ ] Span { { - 2 , 4 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 0 , 0 , 5 , 3 , - 4 , - 2 , 3 , 4 , - 9 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 5 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 4 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , 3 , 4 } ,
} ,
} ,
{
"cut empty buckets at start or end of spans, even in the middle" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 6 } , { 3 , 6 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , 3 , - 4 , 0 , 1 , 42 , 3 , - 46 , 0 , 0 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 2 , 6 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 8 , 4 , - 2 , 3 , 4 , - 9 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 2 } , { 5 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 4 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , 3 , 4 } ,
} ,
} ,
{
"cut empty buckets at start or end but merge spans due to maxEmptyBuckets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 4 } , { 5 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , 3 , - 3 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , 3 , 4 , - 9 } ,
} ,
10 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 10 } } ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 4 , 0 , 0 , 0 , 0 , 1 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 9 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 8 , 0 , 0 , 4 , - 2 , 3 , 4 } ,
} ,
} ,
{
"cut empty buckets from the middle of a span" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 6 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , - 1 , 0 , 3 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , - 2 , 3 , 4 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 1 } , { 2 , 1 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 2 } , { 1 , 2 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , 1 , 4 } ,
} ,
} ,
{
"cut out a span containing only empty buckets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 3 } , { 2 , 2 } , { 3 , 4 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , - 1 , 0 , 3 , - 2 , 42 , 3 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 1 } , { 7 , 4 } } ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 42 , 3 } ,
} ,
} ,
{
"cut empty buckets from the middle of a span, avoiding some due to maxEmptyBuckets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 6 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , - 1 , 0 , 3 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , - 2 , 3 , 4 } ,
} ,
1 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 1 } , { 2 , 1 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , - 2 , 3 , 4 } ,
} ,
} ,
{
"avoiding all cutting of empty buckets from the middle of a chunk due to maxEmptyBuckets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 6 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , - 1 , 0 , 3 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , - 2 , 3 , 4 } ,
} ,
2 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 4 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 1 , - 1 , 0 , 3 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , - 2 , 3 , 4 } ,
} ,
} ,
{
"everything merged into one span due to maxEmptyBuckets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 6 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 1 , - 1 , 0 , 3 , - 2 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 2 } , { 3 , 5 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 4 , - 2 , - 2 , 3 , 4 } ,
} ,
3 ,
& Histogram {
PositiveSpans : [ ] Span { { - 2 , 10 } } ,
PositiveBuckets : [ ] int64 { 1 , - 1 , 0 , 3 , - 3 , 0 , 0 , 1 , 42 , 3 } ,
NegativeSpans : [ ] Span { { 0 , 10 } } ,
NegativeBuckets : [ ] int64 { 5 , 3 , - 8 , 0 , 0 , 4 , - 2 , - 2 , 3 , 4 } ,
} ,
} ,
{
"only empty buckets and maxEmptyBuckets greater zero" ,
& Histogram {
PositiveSpans : [ ] Span { { - 4 , 6 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 , 0 } ,
NegativeSpans : [ ] Span { { 0 , 7 } } ,
NegativeBuckets : [ ] int64 { 0 , 0 , 0 , 0 , 0 , 0 , 0 } ,
} ,
3 ,
& Histogram {
PositiveSpans : [ ] Span { } ,
PositiveBuckets : [ ] int64 { } ,
NegativeSpans : [ ] Span { } ,
NegativeBuckets : [ ] int64 { } ,
} ,
} ,
{
"multiple spans of only empty buckets" ,
& Histogram {
PositiveSpans : [ ] Span { { - 10 , 2 } , { 2 , 1 } , { 3 , 3 } } ,
PositiveBuckets : [ ] int64 { 0 , 0 , 0 , 0 , 2 , 3 } ,
NegativeSpans : [ ] Span { { - 10 , 2 } , { 2 , 1 } , { 3 , 3 } } ,
NegativeBuckets : [ ] int64 { 2 , 3 , - 5 , 0 , 0 , 0 } ,
} ,
0 ,
& Histogram {
PositiveSpans : [ ] Span { { - 1 , 2 } } ,
PositiveBuckets : [ ] int64 { 2 , 3 } ,
NegativeSpans : [ ] Span { { - 10 , 2 } } ,
NegativeBuckets : [ ] int64 { 2 , 3 } ,
} ,
} ,
}
for _ , c := range cases {
t . Run ( c . name , func ( t * testing . T ) {
require . Equal ( t , c . expected , c . in . Compact ( c . maxEmptyBuckets ) )
// Compact has happened in-place, too.
require . Equal ( t , c . expected , c . in )
} )
}
}
func TestHistogramValidation ( t * testing . T ) {
tests := map [ string ] struct {
h * Histogram
errMsg string
skipFloat bool
} {
"valid histogram" : {
h : & Histogram {
Count : 12 ,
ZeroCount : 2 ,
ZeroThreshold : 0.001 ,
Sum : 19.4 ,
Schema : 1 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 2 } ,
{ Offset : 1 , Length : 2 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 1 , - 1 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 2 } ,
{ Offset : 1 , Length : 2 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 1 , - 1 , 0 } ,
} ,
} ,
"valid histogram with NaN observations that has its Count (4) higher than the actual total of buckets (2 + 1)" : {
// This case is possible if NaN values (which do not fall into any bucket) are observed.
h : & Histogram {
ZeroCount : 2 ,
Count : 4 ,
Sum : math . NaN ( ) ,
PositiveSpans : [ ] Span { { Offset : 0 , Length : 1 } } ,
PositiveBuckets : [ ] int64 { 1 } ,
} ,
} ,
"rejects histogram without NaN observations that has its Count (4) higher than the actual total of buckets (2 + 1)" : {
h : & Histogram {
ZeroCount : 2 ,
Count : 4 ,
Sum : 333 ,
PositiveSpans : [ ] Span { { Offset : 0 , Length : 1 } } ,
PositiveBuckets : [ ] int64 { 1 } ,
} ,
errMsg : ` 3 observations found in buckets, but the Count field is 4: histogram's observation count should equal the number of observations found in the buckets (in absence of NaN) ` ,
skipFloat : true ,
} ,
"rejects histogram that has too few negative buckets" : {
h : & Histogram {
NegativeSpans : [ ] Span { { Offset : 0 , Length : 1 } } ,
NegativeBuckets : [ ] int64 { } ,
} ,
errMsg : ` negative side: spans need 1 buckets, have 0 buckets: histogram spans specify different number of buckets than provided ` ,
} ,
"rejects histogram that has too few positive buckets" : {
h : & Histogram {
PositiveSpans : [ ] Span { { Offset : 0 , Length : 1 } } ,
PositiveBuckets : [ ] int64 { } ,
} ,
errMsg : ` positive side: spans need 1 buckets, have 0 buckets: histogram spans specify different number of buckets than provided ` ,
} ,
"rejects histogram that has too many negative buckets" : {
h : & Histogram {
NegativeSpans : [ ] Span { { Offset : 0 , Length : 1 } } ,
NegativeBuckets : [ ] int64 { 1 , 2 } ,
} ,
errMsg : ` negative side: spans need 1 buckets, have 2 buckets: histogram spans specify different number of buckets than provided ` ,
} ,
"rejects histogram that has too many positive buckets" : {
h : & Histogram {
PositiveSpans : [ ] Span { { Offset : 0 , Length : 1 } } ,
PositiveBuckets : [ ] int64 { 1 , 2 } ,
} ,
errMsg : ` positive side: spans need 1 buckets, have 2 buckets: histogram spans specify different number of buckets than provided ` ,
} ,
"rejects a histogram that has a negative span with a negative offset" : {
h : & Histogram {
NegativeSpans : [ ] Span { { Offset : - 1 , Length : 1 } , { Offset : - 1 , Length : 1 } } ,
NegativeBuckets : [ ] int64 { 1 , 2 } ,
} ,
errMsg : ` negative side: span number 2 with offset -1: histogram has a span whose offset is negative ` ,
} ,
"rejects a histogram which has a positive span with a negative offset" : {
h : & Histogram {
PositiveSpans : [ ] Span { { Offset : - 1 , Length : 1 } , { Offset : - 1 , Length : 1 } } ,
PositiveBuckets : [ ] int64 { 1 , 2 } ,
} ,
errMsg : ` positive side: span number 2 with offset -1: histogram has a span whose offset is negative ` ,
} ,
"rejects a histogram that has a negative bucket with a negative count" : {
h : & Histogram {
NegativeSpans : [ ] Span { { Offset : - 1 , Length : 1 } } ,
NegativeBuckets : [ ] int64 { - 1 } ,
} ,
errMsg : ` negative side: bucket number 1 has observation count of -1: histogram has a bucket whose observation count is negative ` ,
} ,
"rejects a histogram that has a positive bucket with a negative count" : {
h : & Histogram {
PositiveSpans : [ ] Span { { Offset : - 1 , Length : 1 } } ,
PositiveBuckets : [ ] int64 { - 1 } ,
} ,
errMsg : ` positive side: bucket number 1 has observation count of -1: histogram has a bucket whose observation count is negative ` ,
} ,
"rejects a histogram that has a lower count than count in buckets" : {
h : & Histogram {
Count : 0 ,
NegativeSpans : [ ] Span { { Offset : - 1 , Length : 1 } } ,
PositiveSpans : [ ] Span { { Offset : - 1 , Length : 1 } } ,
NegativeBuckets : [ ] int64 { 1 } ,
PositiveBuckets : [ ] int64 { 1 } ,
} ,
errMsg : ` 2 observations found in buckets, but the Count field is 0: histogram's observation count should equal the number of observations found in the buckets (in absence of NaN) ` ,
skipFloat : true ,
} ,
"rejects a histogram that doesn't count the zero bucket in its count" : {
h : & Histogram {
Count : 2 ,
ZeroCount : 1 ,
NegativeSpans : [ ] Span { { Offset : - 1 , Length : 1 } } ,
PositiveSpans : [ ] Span { { Offset : - 1 , Length : 1 } } ,
NegativeBuckets : [ ] int64 { 1 } ,
PositiveBuckets : [ ] int64 { 1 } ,
} ,
errMsg : ` 3 observations found in buckets, but the Count field is 2: histogram's observation count should equal the number of observations found in the buckets (in absence of NaN) ` ,
skipFloat : true ,
} ,
}
for testName , tc := range tests {
t . Run ( testName , func ( t * testing . T ) {
if err := tc . h . Validate ( ) ; tc . errMsg != "" {
require . EqualError ( t , err , tc . errMsg )
} else {
require . NoError ( t , err )
}
if tc . skipFloat {
return
}
fh := tc . h . ToFloat ( )
if err := fh . Validate ( ) ; tc . errMsg != "" {
require . EqualError ( t , err , tc . errMsg )
} else {
require . NoError ( t , err )
}
} )
}
}
func BenchmarkHistogramValidation ( b * testing . B ) {
histograms := GenerateBigTestHistograms ( b . N , 500 )
b . ResetTimer ( )
for _ , h := range histograms {
require . NoError ( b , h . Validate ( ) )
}
}
func TestHistogramReduceResolution ( t * testing . T ) {
tcs := map [ string ] struct {
origin * Histogram
target * Histogram
} {
"valid histogram" : {
origin : & Histogram {
Schema : 0 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 0 , Length : 0 } ,
{ Offset : 3 , Length : 2 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 4 } ,
{ Offset : 0 , Length : 0 } ,
{ Offset : 3 , Length : 2 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 2 , - 2 , 1 , - 1 , 0 } ,
} ,
target : & Histogram {
Schema : - 1 ,
PositiveSpans : [ ] Span {
{ Offset : 0 , Length : 3 } ,
{ Offset : 1 , Length : 1 } ,
} ,
PositiveBuckets : [ ] int64 { 1 , 3 , - 2 , 0 } ,
NegativeSpans : [ ] Span {
{ Offset : 0 , Length : 3 } ,
{ Offset : 1 , Length : 1 } ,
} ,
NegativeBuckets : [ ] int64 { 1 , 3 , - 2 , 0 } ,
} ,
} ,
}
for _ , tc := range tcs {
target := tc . origin . ReduceResolution ( tc . target . Schema )
require . Equal ( t , tc . target , target )
// Check that receiver histogram was mutated:
require . Equal ( t , tc . target , tc . origin )
}
}