Documentation
¶
Index ¶
- func BuildDirtyIndex(ctx context.Context, localPath, dirtyPath string, opts IndexOptions) (int, error)
- func CloneOrFetch(url, repoPath string) (*git.Repository, error)
- func DirtyIndexPath(codedbDir, worktreePath string) string
- func IndexLocalRepo(ctx context.Context, s *store.Store, localPath string, opts IndexOptions) error
- func IndexRepo(ctx context.Context, s *store.Store, url string, opts IndexOptions) error
- func RepoDirFromURL(url string) (string, error)
- func RepoNameFromURL(url string) (string, error)
- type BleveCodeDoc
- type BleveCommentDoc
- type BleveDiffDoc
- type CommentStats
- type GitHubIndexStats
- type IndexOptions
- type ParseStats
- type ProgressFunc
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func BuildDirtyIndex ¶ added in v0.5.0
func BuildDirtyIndex(ctx context.Context, localPath, dirtyPath string, opts IndexOptions) (int, error)
BuildDirtyIndex creates an on-disk Bleve index of dirty (uncommitted) files. Uses git status to identify dirty files (fast) rather than walking the full tree. The dirty index at dirtyPath is fully rebuilt on each call via atomic swap (write to .tmp, then rename) so CLI readers never see a partially-written index. Files are indexed as "dirty_<relPath>" for enrichment without SQL.
func CloneOrFetch ¶
func CloneOrFetch(url, repoPath string) (*git.Repository, error)
CloneOrFetch clones a bare repo at repoPath, or fetches if it already exists. Returns the opened go-git Repository.
func DirtyIndexPath ¶ added in v0.5.0
DirtyIndexPath returns the on-disk path for a worktree's dirty overlay index. The path is deterministic based on the worktree's absolute path, allowing the CLI to find the daemon-built dirty index without IPC.
func IndexLocalRepo ¶
IndexLocalRepo indexes a local git repository in-place (no clone). It indexes all committed history AND the current working tree, including uncommitted (dirty) files.
func RepoDirFromURL ¶
RepoDirFromURL derives a local directory name from a repo URL. "https://github.com/user/repo/" -> "github.com/user/repo.git"
func RepoNameFromURL ¶
RepoNameFromURL derives a human-readable name from a URL. "https://github.com/user/repo" -> "github.com/user/repo"
Types ¶
type BleveCodeDoc ¶
type BleveCodeDoc struct {
Content string `json:"content"`
}
BleveCodeDoc is the document indexed into the code Bleve index.
type BleveCommentDoc ¶ added in v0.5.0
type BleveCommentDoc struct {
Content string `json:"content"`
}
BleveCommentDoc is the document indexed into the comment Bleve index.
type BleveDiffDoc ¶
type BleveDiffDoc struct {
Content string `json:"content"`
}
BleveDiffDoc is the document indexed into the diff Bleve index.
type CommentStats ¶ added in v0.5.0
CommentStats holds statistics from the comment parsing phase.
func ParseComments ¶ added in v0.5.0
func ParseComments(ctx context.Context, s *store.Store, progress ProgressFunc) (CommentStats, error)
ParseComments extracts comments from all blobs that haven't been comment-parsed yet and inserts them into the comments table and CommentIndex.
type GitHubIndexStats ¶ added in v0.5.0
GitHubIndexStats tracks how many PRs and issues were indexed.
func IndexGitHubData ¶ added in v0.5.0
func IndexGitHubData(ctx context.Context, s *store.Store, ledgerPath string, progress ProgressFunc) (*GitHubIndexStats, error)
IndexGitHubData reads PR and issue JSON files from the ledger and upserts them into CodeDB's pull_requests/issues tables with their comments.
The ledger stores data in time-partitioned directories:
data/github/YYYY/MM/DD/pr/NNN.json data/github/YYYY/MM/DD/issue/NNN.json
Each file is a self-contained JSON blob that we upsert by number. Existing records are replaced (delete + insert) to pick up state changes, new comments, etc.
Incremental: files are skipped if their mtime hasn't changed since the last successful index. This reduces steady-state cost from O(all files) to O(changed files) per run.
type IndexOptions ¶
type IndexOptions struct {
MaxHistoryDepth int // 0 = unlimited
Progress ProgressFunc
SkipDirs map[string]bool // directories to skip in worktree indexing; nil = use defaultSkipDirs
}
IndexOptions configures the indexing process.
type ParseStats ¶
ParseStats holds statistics from the symbol parsing phase.
func ParseSymbols ¶
func ParseSymbols(ctx context.Context, s *store.Store, progress ProgressFunc) (ParseStats, error)
ParseSymbols extracts symbols and references from all unparsed blobs with supported languages and inserts them into the symbols and symbol_refs tables.
type ProgressFunc ¶
type ProgressFunc func(msg string)
ProgressFunc is called with status messages during indexing.
Source Files