1
1
mirror of https://github.com/go-gitea/gitea synced 2024-11-05 09:44:26 +00:00
gitea/modules/git/repo_language_stats_nogogit.go
flynnnnnnnnnn e81ccc406b
Implement FSFE REUSE for golang files (#21840)
Change all license headers to comply with REUSE specification.

Fix #16132

Co-authored-by: flynnnnnnnnnn <flynnnnnnnnnn@github>
Co-authored-by: John Olheiser <john.olheiser@gmail.com>
2022-11-27 18:20:29 +00:00

233 lines
6.1 KiB
Go

// Copyright 2020 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
//go:build !gogit
package git
import (
"bufio"
"bytes"
"io"
"math"
"strings"
"code.gitea.io/gitea/modules/analyze"
"code.gitea.io/gitea/modules/log"
"github.com/go-enry/go-enry/v2"
)
// GetLanguageStats calculates language stats for git repository at specified commit
func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, error) {
// We will feed the commit IDs in order into cat-file --batch, followed by blobs as necessary.
// so let's create a batch stdin and stdout
batchStdinWriter, batchReader, cancel := repo.CatFileBatch(repo.Ctx)
defer cancel()
writeID := func(id string) error {
_, err := batchStdinWriter.Write([]byte(id + "\n"))
return err
}
if err := writeID(commitID); err != nil {
return nil, err
}
shaBytes, typ, size, err := ReadBatchLine(batchReader)
if typ != "commit" {
log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
return nil, ErrNotExist{commitID, ""}
}
sha, err := NewIDFromString(string(shaBytes))
if err != nil {
log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
return nil, ErrNotExist{commitID, ""}
}
commit, err := CommitFromReader(repo, sha, io.LimitReader(batchReader, size))
if err != nil {
log.Debug("Unable to get commit for: %s. Err: %v", commitID, err)
return nil, err
}
if _, err = batchReader.Discard(1); err != nil {
return nil, err
}
tree := commit.Tree
entries, err := tree.ListEntriesRecursiveWithSize()
if err != nil {
return nil, err
}
checker, deferable := repo.CheckAttributeReader(commitID)
defer deferable()
contentBuf := bytes.Buffer{}
var content []byte
// sizes contains the current calculated size of all files by language
sizes := make(map[string]int64)
// by default we will only count the sizes of programming languages or markup languages
// unless they are explicitly set using linguist-language
includedLanguage := map[string]bool{}
// or if there's only one language in the repository
firstExcludedLanguage := ""
firstExcludedLanguageSize := int64(0)
for _, f := range entries {
select {
case <-repo.Ctx.Done():
return sizes, repo.Ctx.Err()
default:
}
contentBuf.Reset()
content = contentBuf.Bytes()
if f.Size() == 0 {
continue
}
notVendored := false
notGenerated := false
if checker != nil {
attrs, err := checker.CheckPath(f.Name())
if err == nil {
if vendored, has := attrs["linguist-vendored"]; has {
if vendored == "set" || vendored == "true" {
continue
}
notVendored = vendored == "false"
}
if generated, has := attrs["linguist-generated"]; has {
if generated == "set" || generated == "true" {
continue
}
notGenerated = generated == "false"
}
if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
// group languages, such as Pug -> HTML; SCSS -> CSS
group := enry.GetLanguageGroup(language)
if len(group) != 0 {
language = group
}
// this language will always be added to the size
sizes[language] += f.Size()
continue
} else if language, has := attrs["gitlab-language"]; has && language != "unspecified" && language != "" {
// strip off a ? if present
if idx := strings.IndexByte(language, '?'); idx >= 0 {
language = language[:idx]
}
if len(language) != 0 {
// group languages, such as Pug -> HTML; SCSS -> CSS
group := enry.GetLanguageGroup(language)
if len(group) != 0 {
language = group
}
// this language will always be added to the size
sizes[language] += f.Size()
continue
}
}
}
}
if (!notVendored && analyze.IsVendor(f.Name())) || enry.IsDotFile(f.Name()) ||
enry.IsDocumentation(f.Name()) || enry.IsConfiguration(f.Name()) {
continue
}
// If content can not be read or file is too big just do detection by filename
if f.Size() <= bigFileSize {
if err := writeID(f.ID.String()); err != nil {
return nil, err
}
_, _, size, err := ReadBatchLine(batchReader)
if err != nil {
log.Debug("Error reading blob: %s Err: %v", f.ID.String(), err)
return nil, err
}
sizeToRead := size
discard := int64(1)
if size > fileSizeLimit {
sizeToRead = fileSizeLimit
discard = size - fileSizeLimit + 1
}
_, err = contentBuf.ReadFrom(io.LimitReader(batchReader, sizeToRead))
if err != nil {
return nil, err
}
content = contentBuf.Bytes()
err = discardFull(batchReader, discard)
if err != nil {
return nil, err
}
}
if !notGenerated && enry.IsGenerated(f.Name(), content) {
continue
}
// FIXME: Why can't we split this and the IsGenerated tests to avoid reading the blob unless absolutely necessary?
// - eg. do the all the detection tests using filename first before reading content.
language := analyze.GetCodeLanguage(f.Name(), content)
if language == enry.OtherLanguage || language == "" {
continue
}
// group languages, such as Pug -> HTML; SCSS -> CSS
group := enry.GetLanguageGroup(language)
if group != "" {
language = group
}
included, checked := includedLanguage[language]
if !checked {
langtype := enry.GetLanguageType(language)
included = langtype == enry.Programming || langtype == enry.Markup
includedLanguage[language] = included
}
if included {
sizes[language] += f.Size()
} else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
firstExcludedLanguage = language
firstExcludedLanguageSize += f.Size()
}
continue
}
// If there are no included languages add the first excluded language
if len(sizes) == 0 && firstExcludedLanguage != "" {
sizes[firstExcludedLanguage] = firstExcludedLanguageSize
}
return sizes, nil
}
func discardFull(rd *bufio.Reader, discard int64) error {
if discard > math.MaxInt32 {
n, err := rd.Discard(math.MaxInt32)
discard -= int64(n)
if err != nil {
return err
}
}
for discard > 0 {
n, err := rd.Discard(int(discard))
discard -= int64(n)
if err != nil {
return err
}
}
return nil
}