feat: add local PyPI repository support

Upload Python wheels and sdists to local PyPI repos. The simple index (PEP 503) is computed on-demand from stored files. - Upload validates .whl/.tar.gz/.zip filenames, parses and normalizes package names per PEP 503, stores under {package}/{filename} - GET /api/v1/local/{name}/simple/ serves root index listing all packages - GET /api/v1/local/{name}/simple/{pkg}/ serves per-package file listing with sha256 hashes for integrity verification - Files are downloadable at /api/v1/local/{name}/{package}/{filename} - Overwrites rejected with 409 Tested e2e: uv build wheel → upload → uv pip install from local repo
2026-06-23 22:04:12 +10:00
parent 1e91a5fb72
commit bb81eafa71
3 changed files with 213 additions and 2 deletions
@@ -115,10 +115,15 @@ func (h *ProxyHandler) handleLocal(w http.ResponseWriter, r *http.Request) {
 		return
 	}

-	if remote.PackageType == models.PackageTerraform {
+	switch remote.PackageType {
+	case models.PackageTerraform:
 		if h.serveTerraformMirror(w, r, remote, path) {
 			return
 		}
+	case models.PackagePyPI:
+		if h.servePyPIMirror(w, r, remote, path) {
+			return
+		}
 	}

 	h.serveLocalFile(w, r, localName, path)
@@ -149,6 +154,24 @@ func (h *ProxyHandler) serveTerraformMirror(w http.ResponseWriter, r *http.Reque
 	return false
 }

+func (h *ProxyHandler) servePyPIMirror(w http.ResponseWriter, r *http.Request, remote *models.Remote, path string) bool {
+	if path == "simple" || path == "simple/" {
+		h.local.ServePyPIIndex(w, r, remote.Name)
+		return true
+	}
+
+	if strings.HasPrefix(path, "simple/") {
+		pkg := strings.TrimPrefix(path, "simple/")
+		pkg = strings.TrimSuffix(pkg, "/")
+		if pkg != "" && !strings.Contains(pkg, "/") {
+			h.local.ServePyPIPackageIndex(w, r, remote.Name, pkg)
+			return true
+		}
+	}
+
+	return false
+}
+
 func (h *ProxyHandler) serveLocalFile(w http.ResponseWriter, r *http.Request, repoName, path string) {
 	file, err := h.db.GetLocalFile(r.Context(), repoName, path)
 	if err != nil {
@@ -1,6 +1,7 @@
 package v2

 import (
+	"context"
 	"encoding/json"
 	"errors"
 	"fmt"
@@ -16,6 +17,37 @@ import (
 	"git.unkin.net/unkin/artifactapi/pkg/models"
 )

+var pypiFileRe = regexp.MustCompile(`^[a-zA-Z0-9][a-zA-Z0-9._-]*\.(whl|tar\.gz|zip)$`)
+
+var pypiNormalizeRe = regexp.MustCompile(`[-_.]+`)
+
+func pypiNormalize(name string) string {
+	return strings.ToLower(pypiNormalizeRe.ReplaceAllString(name, "-"))
+}
+
+func pypiPackageFromWheel(filename string) string {
+	parts := strings.SplitN(filename, "-", 3)
+	if len(parts) < 2 {
+		return ""
+	}
+	return pypiNormalize(parts[0])
+}
+
+func pypiPackageFromSdist(filename string) string {
+	name := filename
+	for _, suffix := range []string{".tar.gz", ".zip"} {
+		if strings.HasSuffix(name, suffix) {
+			name = strings.TrimSuffix(name, suffix)
+			break
+		}
+	}
+	idx := strings.LastIndex(name, "-")
+	if idx <= 0 {
+		return ""
+	}
+	return pypiNormalize(name[:idx])
+}
+
 var providerZipRe = regexp.MustCompile(
 	`^terraform-provider-([a-zA-Z0-9_-]+)_([0-9]+\.[0-9]+\.[0-9]+(?:-[a-zA-Z0-9.]+)?)_([a-z0-9]+)_([a-z0-9]+)\.zip$`,
 )
@@ -61,9 +93,13 @@ func (h *LocalHandler) upload(w http.ResponseWriter, r *http.Request) {
 		return
 	}

-	if remote.PackageType == models.PackageTerraform {
+	switch remote.PackageType {
+	case models.PackageTerraform:
 		h.uploadTerraformProvider(w, r, remote, filePath)
 		return
+	case models.PackagePyPI:
+		h.uploadPyPI(w, r, remote, filePath)
+		return
 	}

 	h.uploadGeneric(w, r, remote, filePath)
@@ -184,6 +220,135 @@ func (h *LocalHandler) uploadGeneric(w http.ResponseWriter, r *http.Request, rem
 	})
 }

+func (h *LocalHandler) uploadPyPI(w http.ResponseWriter, r *http.Request, remote *models.Remote, filePath string) {
+	filename := filePath
+	if idx := strings.LastIndex(filePath, "/"); idx >= 0 {
+		filename = filePath[idx+1:]
+	}
+
+	if !pypiFileRe.MatchString(filename) {
+		http.Error(w, fmt.Sprintf("filename %q must be a .whl, .tar.gz, or .zip file", filename), http.StatusBadRequest)
+		return
+	}
+
+	var pkgName string
+	if strings.HasSuffix(filename, ".whl") {
+		pkgName = pypiPackageFromWheel(filename)
+	} else {
+		pkgName = pypiPackageFromSdist(filename)
+	}
+	if pkgName == "" {
+		http.Error(w, fmt.Sprintf("cannot parse package name from %q", filename), http.StatusBadRequest)
+		return
+	}
+
+	storagePath := pkgName + "/" + filename
+
+	existing, err := h.db.GetLocalFile(r.Context(), remote.Name, storagePath)
+	if err != nil {
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+	if existing != nil {
+		http.Error(w, fmt.Sprintf("file %q already exists; overwrites are not allowed", storagePath), http.StatusConflict)
+		return
+	}
+
+	contentType := "application/zip"
+	if strings.HasSuffix(filename, ".tar.gz") {
+		contentType = "application/gzip"
+	}
+
+	result, err := h.cas.Store(r.Context(), r.Body, contentType)
+	if err != nil {
+		http.Error(w, fmt.Sprintf("store failed: %v", err), http.StatusInternalServerError)
+		return
+	}
+
+	if err := h.db.UpsertBlob(r.Context(), result.ContentHash, result.S3Key, result.SizeBytes, contentType); err != nil {
+		http.Error(w, fmt.Sprintf("record blob: %v", err), http.StatusInternalServerError)
+		return
+	}
+
+	if err := h.db.CreateLocalFile(r.Context(), remote.Name, storagePath, result.ContentHash); err != nil {
+		if errors.Is(err, database.ErrAlreadyExists) {
+			http.Error(w, fmt.Sprintf("file %q already exists; overwrites are not allowed", storagePath), http.StatusConflict)
+			return
+		}
+		http.Error(w, fmt.Sprintf("record file: %v", err), http.StatusInternalServerError)
+		return
+	}
+
+	writeJSON(w, http.StatusCreated, map[string]any{
+		"package":      pkgName,
+		"filename":     filename,
+		"content_hash": result.ContentHash,
+		"size_bytes":   result.SizeBytes,
+	})
+}
+
+func (h *LocalHandler) ServePyPIIndex(w http.ResponseWriter, r *http.Request, repoName string) {
+	packages, err := h.db.ListLocalFilePackages(r.Context(), repoName)
+	if err != nil {
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	var b strings.Builder
+	b.WriteString("<!DOCTYPE html>\n<html><body>\n")
+	for _, pkg := range packages {
+		fmt.Fprintf(&b, "<a href=\"%s/\">%s</a>\n", pkg, pkg)
+	}
+	b.WriteString("</body></html>\n")
+
+	w.Header().Set("Content-Type", "text/html")
+	w.WriteHeader(http.StatusOK)
+	io.WriteString(w, b.String())
+}
+
+func (h *LocalHandler) ServePyPIPackageIndex(w http.ResponseWriter, r *http.Request, repoName, packageName string) {
+	normalized := pypiNormalize(packageName)
+	prefix := normalized + "/"
+	files, err := h.db.ListLocalFilesByPrefix(r.Context(), repoName, prefix)
+	if err != nil {
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	if len(files) == 0 {
+		http.Error(w, "not found", http.StatusNotFound)
+		return
+	}
+
+	body := h.generatePyPIPackageHTML(normalized, files)
+	w.Header().Set("Content-Type", "text/html")
+	w.WriteHeader(http.StatusOK)
+	w.Write(body)
+}
+
+func (h *LocalHandler) GeneratePyPIPackageHTML(ctx context.Context, repoName, packageName string) ([]byte, error) {
+	normalized := pypiNormalize(packageName)
+	prefix := normalized + "/"
+	files, err := h.db.ListLocalFilesByPrefix(ctx, repoName, prefix)
+	if err != nil {
+		return nil, err
+	}
+	return h.generatePyPIPackageHTML(normalized, files), nil
+}
+
+func (h *LocalHandler) generatePyPIPackageHTML(packageName string, files []database.LocalFile) []byte {
+	var b strings.Builder
+	b.WriteString("<!DOCTYPE html>\n<html><body>\n")
+	for _, f := range files {
+		filename := strings.TrimPrefix(f.FilePath, packageName+"/")
+		hash := strings.TrimPrefix(f.ContentHash, "sha256:")
+		fmt.Fprintf(&b, "<a href=\"../../%s/%s#sha256=%s\">%s</a>\n",
+			packageName, filename, hash, filename)
+	}
+	b.WriteString("</body></html>\n")
+	return []byte(b.String())
+}
+
 func (h *LocalHandler) download(w http.ResponseWriter, r *http.Request) {
 	repoName := chi.URLParam(r, "name")
 	filePath := chi.URLParam(r, "*")
@@ -99,6 +99,29 @@ func (db *DB) ListLocalFilesByPrefix(ctx context.Context, repoName, prefix strin
 	return files, rows.Err()
 }

+func (db *DB) ListLocalFilePackages(ctx context.Context, repoName string) ([]string, error) {
+	rows, err := db.Pool.Query(ctx, `
+		SELECT DISTINCT split_part(file_path, '/', 1)
+		FROM local_files
+		WHERE repo_name = $1
+		ORDER BY 1
+	`, repoName)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+
+	var packages []string
+	for rows.Next() {
+		var pkg string
+		if err := rows.Scan(&pkg); err != nil {
+			return nil, err
+		}
+		packages = append(packages, pkg)
+	}
+	return packages, rows.Err()
+}
+
 func (db *DB) DeleteLocalFile(ctx context.Context, repoName, filePath string) error {
 	_, err := db.Pool.Exec(ctx, `DELETE FROM local_files WHERE repo_name = $1 AND file_path = $2`, repoName, filePath)
 	return err