update QoL and ai integration

2026-07-29 14:16:14 +02:00
parent 1d90e364ee
commit 027057e85f
14 changed files with 1893 additions and 84 deletions
--- a/ai_diff.go
+++ b/ai_diff.go
@@ -3,6 +3,8 @@ package main
 import (
 	"bufio"
 	"context"
+	"encoding/base64"
+	"encoding/json"
 	"fmt"
 	"io"
 	"net/http"
@@ -50,6 +52,149 @@ func (c *CachedGitHubService) PullRequestDiff(ctx context.Context, owner, repo s
 	return service.PullRequestDiff(ctx, owner, repo, number)
 }

+func (c *GitHubClient) RepositoryTree(
+	ctx context.Context, owner, repo, commitOID string,
+) (AIRepositoryTree, error) {
+	baseURL := c.restBaseURL() + "/repos/" + url.PathEscape(owner) + "/" + url.PathEscape(repo)
+	var commit struct {
+		Tree struct {
+			SHA string `json:"sha"`
+		} `json:"tree"`
+	}
+	if err := c.getAIRepositoryJSON(
+		ctx, baseURL+"/git/commits/"+url.PathEscape(commitOID), 1<<20, &commit,
+	); err != nil {
+		return AIRepositoryTree{}, fmt.Errorf("load repository commit: %w", err)
+	}
+	if commit.Tree.SHA == "" {
+		return AIRepositoryTree{}, fmt.Errorf("load repository commit: GitHub returned no tree OID")
+	}
+	var tree struct {
+		Truncated bool `json:"truncated"`
+		Tree      []struct {
+			Path string `json:"path"`
+			Mode string `json:"mode"`
+			Type string `json:"type"`
+			SHA  string `json:"sha"`
+			Size int64  `json:"size"`
+		} `json:"tree"`
+	}
+	if err := c.getAIRepositoryJSON(
+		ctx, baseURL+"/git/trees/"+url.PathEscape(commit.Tree.SHA)+"?recursive=1",
+		8<<20, &tree,
+	); err != nil {
+		return AIRepositoryTree{}, fmt.Errorf("load repository tree: %w", err)
+	}
+	result := AIRepositoryTree{
+		CommitOID: commitOID,
+		Entries:   make([]AIRepositoryEntry, 0, len(tree.Tree)),
+		Truncated: tree.Truncated,
+	}
+	for _, entry := range tree.Tree {
+		clean := filepath.ToSlash(filepath.Clean(entry.Path))
+		if clean == "." || clean == "" || filepath.IsAbs(clean) ||
+			strings.HasPrefix(clean, "../") {
+			continue
+		}
+		result.Entries = append(result.Entries, AIRepositoryEntry{
+			Path: clean, OID: entry.SHA, Type: entry.Type, Mode: entry.Mode, Size: entry.Size,
+		})
+	}
+	return result, nil
+}
+
+func (c *CachedGitHubService) RepositoryTree(
+	ctx context.Context, owner, repo, commitOID string,
+) (AIRepositoryTree, error) {
+	service, ok := c.remote.(AIRepositoryService)
+	if !ok {
+		return AIRepositoryTree{}, fmt.Errorf("configured GitHub service cannot load repository trees")
+	}
+	return service.RepositoryTree(ctx, owner, repo, commitOID)
+}
+
+func (c *GitHubClient) RepositoryBlob(
+	ctx context.Context, owner, repo, blobOID string, maxBytes int,
+) ([]byte, error) {
+	baseURL := c.restBaseURL() + "/repos/" + url.PathEscape(owner) + "/" + url.PathEscape(repo)
+	var blob struct {
+		Content  string `json:"content"`
+		Encoding string `json:"encoding"`
+		Size     int    `json:"size"`
+		SHA      string `json:"sha"`
+	}
+	responseLimit := int64(max(16_384, maxBytes*2+8_192))
+	if err := c.getAIRepositoryJSON(
+		ctx, baseURL+"/git/blobs/"+url.PathEscape(blobOID), responseLimit, &blob,
+	); err != nil {
+		return nil, fmt.Errorf("load repository blob: %w", err)
+	}
+	if blob.SHA != "" && blob.SHA != blobOID {
+		return nil, fmt.Errorf("load repository blob: GitHub returned an unexpected blob OID")
+	}
+	if blob.Size > maxBytes {
+		return nil, fmt.Errorf("repository file exceeds the %d-byte AI limit", maxBytes)
+	}
+	if blob.Encoding != "base64" {
+		return nil, fmt.Errorf("load repository blob: unsupported encoding %q", blob.Encoding)
+	}
+	content, err := base64.StdEncoding.DecodeString(strings.Map(func(r rune) rune {
+		if r == '\r' || r == '\n' || r == ' ' || r == '\t' {
+			return -1
+		}
+		return r
+	}, blob.Content))
+	if err != nil {
+		return nil, fmt.Errorf("decode repository blob: %w", err)
+	}
+	if len(content) > maxBytes {
+		return nil, fmt.Errorf("repository file exceeds the %d-byte AI limit", maxBytes)
+	}
+	return content, nil
+}
+
+func (c *CachedGitHubService) RepositoryBlob(
+	ctx context.Context, owner, repo, blobOID string, maxBytes int,
+) ([]byte, error) {
+	service, ok := c.remote.(AIRepositoryService)
+	if !ok {
+		return nil, fmt.Errorf("configured GitHub service cannot load repository blobs")
+	}
+	return service.RepositoryBlob(ctx, owner, repo, blobOID, maxBytes)
+}
+
+func (c *GitHubClient) getAIRepositoryJSON(
+	ctx context.Context, requestURL string, limit int64, output any,
+) error {
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, requestURL, nil)
+	if err != nil {
+		return err
+	}
+	req.Header.Set("Authorization", "Bearer "+c.token)
+	req.Header.Set("Accept", "application/vnd.github+json")
+	req.Header.Set("User-Agent", "diple")
+	response, err := c.http.Do(req)
+	if err != nil {
+		return err
+	}
+	defer response.Body.Close()
+	if response.StatusCode != http.StatusOK {
+		body, _ := io.ReadAll(io.LimitReader(response.Body, 4096))
+		return fmt.Errorf("GitHub returned %s: %s", response.Status, strings.TrimSpace(string(body)))
+	}
+	data, err := io.ReadAll(io.LimitReader(response.Body, limit+1))
+	if err != nil {
+		return err
+	}
+	if int64(len(data)) > limit {
+		return fmt.Errorf("GitHub response exceeds the %d-byte safety limit", limit)
+	}
+	if err := json.Unmarshal(data, output); err != nil {
+		return fmt.Errorf("decode GitHub response: %w", err)
+	}
+	return nil
+}
+
 func prepareAIDiff(raw string, config AIConfig) ([]aiDiffFile, []string, int) {
 	sections := strings.Split(raw, "\ndiff --git ")
 	var files []aiDiffFile
@@ -159,21 +304,31 @@ func aiExcludedReason(file aiDiffFile, config AIConfig) string {
 		strings.Contains(file.Text, "Binary files ") || strings.IndexByte(file.Text, 0) >= 0 {
 		return "binary"
 	}
-	lower := strings.ToLower(file.Path)
-	for _, pattern := range config.Exclude {
+	if aiPathMatches(file.Path, config.SensitivePaths) {
+		return "sensitive"
+	}
+	if aiPathMatches(file.Path, config.Exclude) {
+		return "excluded"
+	}
+	return ""
+}
+
+func aiPathMatches(filePath string, patterns []string) bool {
+	lower := strings.ToLower(filepath.ToSlash(filePath))
+	for _, pattern := range patterns {
 		pattern = filepath.ToSlash(pattern)
 		if strings.HasSuffix(pattern, "/") {
 			directory := strings.ToLower(pattern)
 			if strings.HasPrefix(lower, directory) || strings.Contains(lower, "/"+directory) {
-				return "excluded"
+				return true
 			}
 		}
 		if matched, _ := path.Match(strings.ToLower(pattern), path.Base(lower)); matched {
-			return "excluded"
+			return true
 		}
 		if matched, _ := path.Match(strings.ToLower(pattern), lower); matched {
-			return "excluded"
+			return true
 		}
 	}
-	return ""
+	return false
 }