Merge pull request #92 from schedawg74/fix-map-exclude-filtering

fix(mapper): apply configured path excludes

Author: steipete

CHANGELOG.md

@@ -27,6 +27,7 @@
 - Improved Flask route mapping to preserve static blueprint URL prefixes, thanks @rohitjavvadi.
 - Improved Django route mapping to preserve literal `include()` route prefixes, thanks @rohitjavvadi.
 - Added conservative Rails route mapping for literal root and HTTP verb routes, thanks @rohitjavvadi.
+- Fixed heuristic feature mapping to honor configured path include/exclude filters, thanks @schedawg74.
 - Fixed Express route mapping for aliased Router imports that follow block comment banners, thanks @rohitjavvadi.
 - Fixed Laravel route mapping to include array-style `Route::group` prefixes, thanks @rohitjavvadi.
 - Fixed Fastify route-object mapping to emit static method arrays while ignoring dynamic entries, thanks @rohitjavvadi.

src/agent-mapper.ts

@@ -7,7 +7,14 @@ import { pathExists } from "./fs.js";
 import { runCommandArgs } from "./exec.js";
 import { mapFeatureSeeds, MapResult } from "./mapper.js";
 import { FeatureSeed, SeedFileRef, SeedTestRef } from "./mappers/types.js";
-import { isSafeFile, normalize, shouldSkip, walk } from "./mappers/shared.js";
+import {
+  applyPathFilters,
+  isSafeFile,
+  normalize,
+  PathFilters,
+  shouldSkip,
+  walk,
+} from "./mappers/shared.js";
 
 type AgentMapMode = "heuristic" | "auto" | "agent";
 
@@ -26,15 +33,10 @@ type AgentMapOptions = {
   source: AgentMapMode;
   provider: Provider | null;
   providerOptions: ProviderOptions;
-  inventory?: InventoryFilters;
+  inventory?: PathFilters;
   onProgress?: (event: string, fields: Record<string, string | number | boolean>) => void;
 };
 
-type InventoryFilters = {
-  include: string[];
-  exclude: string[];
-};
-
 type RepoInventorySummary = {
   files: number;
   sourceFiles: number;
@@ -148,6 +150,7 @@ export async function mapWithSource(
     options.provider,
     options.providerOptions,
     inventory,
+    options.inventory,
   );
   options.onProgress?.("agent-done", {
     features: agent.features.length,
@@ -199,6 +202,7 @@ async function agentMap(
   provider: Provider,
   providerOptions: ProviderOptions,
   inventory: RepoInventory,
+  filters: PathFilters | undefined,
 ): Promise<MapResult> {
   const prompt = buildAgentMapPrompt(project, {
     manifests: inventory.manifests,
@@ -212,12 +216,10 @@ async function agentMap(
   const seeds = await Promise.all(
     output.features.map((feature) => toSeed(root, feature, inventory.allFiles)),
   );
-  return mapFeatureSeeds(
-    root,
-    project,
-    existing,
-    uniqueSeeds(seeds.filter((seed): seed is FeatureSeed => seed !== null)),
-  );
+  const mappedSeeds = uniqueSeeds(seeds.filter((seed): seed is FeatureSeed => seed !== null));
+  return filters === undefined
+    ? mapFeatureSeeds(root, project, existing, mappedSeeds)
+    : mapFeatureSeeds(root, project, existing, mappedSeeds, { filters });
 }
 
 async function toSeed(
@@ -385,10 +387,10 @@ async function repoInventory(
   root: string,
   project: ProjectRecord,
   features: FeatureRecord[],
-  filters: InventoryFilters | undefined,
+  filters: PathFilters | undefined,
 ): Promise<RepoInventory> {
   const skipPath = await inventorySkipPath(root, project, features);
-  const files = applyInventoryFilters(
+  const files = applyPathFilters(
     ((await gitInventoryFiles(root)) ?? (await walk(root, [""], skipPath))).filter(
       (path) => !skipPath(path),
     ),
@@ -443,73 +445,12 @@ async function gitInventoryFiles(root: string): Promise<string[] | null> {
   return existing.filter((path): path is string => path !== null);
 }
 
-function applyInventoryFilters(files: string[], filters: InventoryFilters | undefined): string[] {
-  if (filters === undefined) {
-    return files;
-  }
-  return files.filter(
-    (file) =>
-      filters.include.some((pattern) => inventoryPatternMatches(pattern, file)) &&
-      !filters.exclude.some((pattern) => inventoryPatternMatches(pattern, file)),
-  );
-}
-
 function isInventoryPath(path: string): boolean {
   return (
     path.length > 0 && !isAbsolute(path) && !path.includes("\0") && !path.split("/").includes("..")
   );
 }
 
-function inventoryPatternMatches(pattern: string, path: string): boolean {
-  const normalized = pattern.replace(/\\/gu, "/").replace(/^\.\//u, "");
-  if (normalized === "**" || normalized === "**/*") {
-    return true;
-  }
-  if (normalized.length === 0) {
-    return false;
-  }
-  if (!/[?*]/u.test(normalized)) {
-    return path === normalized || path.startsWith(`${normalized}/`);
-  }
-  if (normalized.endsWith("/**")) {
-    const prefix = normalized.slice(0, -3);
-    if (/[?*]/u.test(prefix)) {
-      return new RegExp(`^${globPatternRegExp(prefix)}(?:/.*)?$`, "u").test(path);
-    }
-    return prefix.length === 0 || path === prefix || path.startsWith(`${prefix}/`);
-  }
-  return new RegExp(`^${globPatternRegExp(normalized)}$`, "u").test(path);
-}
-
-function globPatternRegExp(pattern: string): string {
-  let source = "";
-  for (let index = 0; index < pattern.length; index += 1) {
-    const char = pattern[index];
-    if (char === "*") {
-      if (pattern[index + 1] === "*") {
-        if (pattern[index + 2] === "/") {
-          source += "(?:.*/)?";
-          index += 2;
-        } else {
-          source += ".*";
-          index += 1;
-        }
-      } else {
-        source += "[^/]*";
-      }
-    } else if (char === "?") {
-      source += "[^/]";
-    } else {
-      source += regexpEscape(char ?? "");
-    }
-  }
-  return source;
-}
-
-function regexpEscape(value: string): string {
-  return value.replace(/[\\^$.*+?()[\]{}|]/gu, "\\$&");
-}
-
 async function inventorySkipPath(
   root: string,
   project: ProjectRecord,

src/app.ts

@@ -123,12 +123,14 @@ export async function mapCommand(
   const config = applyProviderFlags(loaded.config, flags);
   const provider = source === "heuristic" ? null : providerByName(config.provider.name);
   const existing = await readFeatures(loaded.paths);
+  const filters = { include: config.include, exclude: config.exclude };
   emitProgress(context, "map", "start", {
     source,
     existing: existing.length,
     dryRun: flags["dryRun"] === true,
   });
   const heuristic = await mapFeatures(loaded.root, loaded.project, existing, {
+    filters,
     onProgress: (event) => {
       emitProgress(context, "map", event.event, {
         mapper: event.mapper,
@@ -149,7 +151,7 @@ export async function mapCommand(
     source,
     provider,
     providerOptions: providerOptions(config),
-    inventory: { include: config.include, exclude: config.exclude },
+    inventory: filters,
     onProgress: (event, fields) => {
       emitProgress(context, "map", event, fields);
     },

src/mapper.test.ts

@@ -10,6 +10,37 @@ import { fixtureRoot, writeFixture } from "./test-helpers.js";
 const symlinkIt = process.platform === "win32" ? it.skip : it;
 
 describe("mapFeatures", () => {
+  it("applies configured path excludes to heuristic feature mapping", async () => {
+    const root = await fixtureRoot("clawpatch-map-exclude-");
+    await writeFixture(root, "requirements.txt", "pytest\n");
+    await writeFixture(root, "src/app/api_service.py", "def call_api(): pass\n");
+    for (let index = 0; index < 13; index += 1) {
+      await writeFixture(
+        root,
+        `src/client/generated/models/model_${index}.py`,
+        `class Model${index}: pass\n`,
+      );
+    }
+
+    const project = await detectProject(root);
+    const result = await mapFeatures(root, project, [], {
+      filters: {
+        include: ["**/*"],
+        exclude: ["src/client/generated/**"],
+      },
+    });
+    const featurePaths = result.features.flatMap((feature) => [
+      ...feature.entrypoints.map((entrypoint) => entrypoint.path),
+      ...feature.ownedFiles.map((file) => file.path),
+      ...feature.contextFiles.map((file) => file.path),
+      ...feature.tests.map((test) => test.path),
+    ]);
+
+    expect(featurePaths).toContain("src/app/api_service.py");
+    expect(result.features.some((feature) => feature.title.includes("generated"))).toBe(false);
+    expect(featurePaths.some((path) => path.startsWith("src/client/generated/"))).toBe(false);
+  });
+
   it("maps package bins, scripts, configs, and Next routes", async () => {
     const root = await fixtureRoot("clawpatch-map-");
     await writeFixture(

src/mapper.ts

@@ -17,7 +17,7 @@ import { reactSeeds } from "./mappers/react.js";
 import { discoverNodeProjects } from "./mappers/projects.js";
 import { rubySeeds } from "./mappers/ruby.js";
 import { rustSeeds } from "./mappers/rust.js";
-import { nearbyTests } from "./mappers/shared.js";
+import { nearbyTests, PathFilters, pathMatchesFilters } from "./mappers/shared.js";
 import { swiftSeeds } from "./mappers/swift.js";
 import { turboTaskGraph } from "./mappers/turbo.js";
 import { FeatureMapper, FeatureSeed, MapperContext } from "./mappers/types.js";
@@ -39,6 +39,7 @@ export type MapProgressEvent = {
 
 export type MapOptions = {
   onProgress?: (event: MapProgressEvent) => void;
+  filters?: PathFilters;
 };
 
 const featureMappers: FeatureMapper[] = [
@@ -68,21 +69,26 @@ export async function mapFeatures(
   options: MapOptions = {},
 ): Promise<MapResult> {
   const seeds = await collectSeeds(root, options);
-  return mapFeatureSeeds(root, project, existing, seeds);
+  return mapFeatureSeeds(root, project, existing, seeds, options);
 }
 
 export async function mapFeatureSeeds(
   root: string,
   project: ProjectRecord,
   existing: FeatureRecord[],
   seeds: FeatureSeed[],
+  options: MapOptions = {},
 ): Promise<MapResult> {
   const existingById = new Map(existing.map((feature) => [feature.featureId, feature]));
   const features: FeatureRecord[] = [];
   let created = 0;
   let changed = 0;
   const now = nowIso();
-  for (const seed of seeds) {
+  for (const rawSeed of seeds) {
+    const seed = filterSeed(rawSeed, options.filters);
+    if (seed === null) {
+      continue;
+    }
     const identity = featureIdentity(seed, existingById);
     const featureId = identity.featureId;
     const previous = existingById.get(featureId);
@@ -100,9 +106,11 @@ export async function mapFeatureSeeds(
               (name): name is string => typeof name === "string",
             ),
           );
-    const tests = uniqueTests([...(seed.tests ?? []), ...discoveredTests]);
+    const tests = uniqueTests(
+      filterTests([...(seed.tests ?? []), ...discoveredTests], options.filters),
+    );
     const contextFiles = uniqueFileRefs([
-      ...(seed.contextFiles ?? []),
+      ...filterFileRefs(seed.contextFiles ?? [], options.filters),
       ...tests.map((test) => ({ path: test.path, reason: "nearby test" })),
     ]);
     const feature: FeatureRecord = {
@@ -159,6 +167,53 @@ export async function mapFeatureSeeds(
   };
 }
 
+function filterSeed(seed: FeatureSeed, filters: PathFilters | undefined): FeatureSeed | null {
+  if (filters === undefined) {
+    return seed;
+  }
+  const ownedFiles =
+    seed.ownedFiles === undefined ? undefined : filterFileRefs(seed.ownedFiles, filters);
+  if (seed.ownedFiles !== undefined && ownedFiles?.length === 0) {
+    return null;
+  }
+  const entryPath = pathMatchesFilters(seed.entryPath, filters)
+    ? seed.entryPath
+    : (ownedFiles?.[0]?.path ?? null);
+  if (entryPath === null) {
+    return null;
+  }
+  const filteredSeed = {
+    ...seed,
+    entryPath,
+    contextFiles: filterFileRefs(seed.contextFiles ?? [], filters),
+    tests: filterTests(seed.tests ?? [], filters),
+  };
+  if (ownedFiles === undefined) {
+    return filteredSeed;
+  }
+  return { ...filteredSeed, ownedFiles };
+}
+
+function filterFileRefs(
+  refs: Array<{ path: string; reason: string }>,
+  filters: PathFilters | undefined,
+): Array<{ path: string; reason: string }> {
+  if (filters === undefined) {
+    return refs;
+  }
+  return refs.filter((ref) => pathMatchesFilters(ref.path, filters));
+}
+
+function filterTests(
+  tests: Array<{ path: string; command: string | null }>,
+  filters: PathFilters | undefined,
+): Array<{ path: string; command: string | null }> {
+  if (filters === undefined) {
+    return tests;
+  }
+  return tests.filter((test) => pathMatchesFilters(test.path, filters));
+}
+
 function featureIdentity(
   seed: FeatureSeed,
   existingById: Map<string, FeatureRecord>,