Multi branch / tag support (#58)

CHANGELOG.md

@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- Added support for indexing and searching repositories across multiple revisions (tag or branch). ([#58](https://github.com/sourcebot-dev/sourcebot/pull/58))
+
 ## [2.3.0] - 2024-11-01
 
 ### Added

README.md

@@ -267,6 +267,46 @@ docker run -e <b>GITEA_TOKEN=my-secret-token</b> /* additional args */ ghcr.io/s
 
 If you're using a self-hosted GitLab or GitHub instance with a custom domain, you can specify the domain in your config file. See [configs/self-hosted.json](configs/self-hosted.json) for examples.
 
+## Searching multiple branches
+
+By default, Sourcebot will index the default branch. To configure Sourcebot to index multiple branches (or tags), the `revisions` field can be used:
+
+```jsonc
+{
+    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "github",
+            "revisions": {
+                // Index the `main` branch and any branches matching the `releases/*` glob pattern.
+                "branches": [
+                    "main",
+                    "releases/*"
+                ],
+                // Index the `latest` tag and any tags matching the `v*.*.*` glob pattern.
+                "tags": [
+                    "latest",
+                    "v*.*.*"
+                ]
+            },
+            "repos": [
+                "my_org/repo_a",
+                "my_org/repo_b"
+            ]
+        }
+    ]
+}
+```
+
+For each repository (in this case, `repo_a` and `repo_b`), Sourcebot will index all branches and tags matching the `branches` and `tags` patterns provided. Any branches or tags that don't match the patterns will be ignored and not indexed.
+
+To search on a specific revision, use the `revision` filter in the search bar:
+
+<picture>
+  <source media="(prefers-color-scheme: dark)" srcset=".github/images/revisions_filter_dark.png">
+  <img style="max-width:700px;width:100%" src=".github/images/revisions_filter_light.png">
+</picture>
+
 ## Searching a local directory
 
 Local directories can be searched by using the `local` type in your config file:

configs/multi-branch.json

@@ -0,0 +1,26 @@
+{
+    "$schema": "../schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "github",
+            "revisions": {
+                 // Specify branches to index...
+                "branches": [
+                    "main",
+                    "release/*"
+                ],
+                // ... or specify tags
+                "tags": [
+                    "v*.*.*"
+                ]
+            },
+            // For each repo (repoa, repob), Sourcebot will index all branches and tags in the repo
+            // matching the `branches` and `tags` patterns above. Any branches or tags that don't
+            // match the patterns will be ignored and not indexed.
+            "repos": [
+                "org/repoa",
+                "org/repob"
+            ]
+        }
+   ]
+}

packages/backend/package.json

@@ -12,6 +12,7 @@
     },
     "devDependencies": {
         "@types/argparse": "^2.0.16",
+        "@types/micromatch": "^4.0.9",
         "@types/node": "^22.7.5",
         "json-schema-to-typescript": "^15.0.2",
         "tsc-watch": "^6.2.0",
@@ -25,6 +26,7 @@
         "cross-fetch": "^4.0.0",
         "gitea-js": "^1.22.0",
         "lowdb": "^7.0.1",
+        "micromatch": "^4.0.8",
         "simple-git": "^3.27.0",
         "strip-json-comments": "^5.0.1",
         "winston": "^3.15.0"

packages/backend/src/gitea.ts

@@ -5,6 +5,7 @@ import { AppContext, GitRepository } from './types.js';
 import fetch from 'cross-fetch';
 import { createLogger } from './logger.js';
 import path from 'path';
+import micromatch from 'micromatch';
 
 const logger = createLogger('Gitea');
 
@@ -60,7 +61,9 @@ export const getGiteaReposFromConfig = async (config: GiteaConfig, ctx: AppConte
                     'zoekt.archived': marshalBool(repo.archived),
                     'zoekt.fork': marshalBool(repo.fork!),
                     'zoekt.public': marshalBool(repo.internal === false && repo.private === false),
-                }
+                },
+                branches: [],
+                tags: []
             } satisfies GitRepository;
         });
     
@@ -78,9 +81,67 @@ export const getGiteaReposFromConfig = async (config: GiteaConfig, ctx: AppConte
         }
     }
 
+    logger.debug(`Found ${repos.length} total repositories.`);
+
+    if (config.revisions) {
+        if (config.revisions.branches) {
+            const branchGlobs = config.revisions.branches;
+            repos = await Promise.all(
+                repos.map(async (repo) => {
+                    const [owner, name] = repo.name.split('/');
+                    let branches = (await getBranchesForRepo(owner, name, api)).map(branch => branch.name!);
+                    branches = micromatch.match(branches, branchGlobs);
+
+                    return {
+                        ...repo,
+                        branches,
+                    };
+                })
+            )
+        }
+
+        if (config.revisions.tags) {
+            const tagGlobs = config.revisions.tags;
+            repos = await Promise.all(
+                repos.map(async (repo) => {
+                    const [owner, name] = repo.name.split('/');
+                    let tags = (await getTagsForRepo(owner, name, api)).map(tag => tag.name!);
+                    tags = micromatch.match(tags, tagGlobs);
+
+                    return {
+                        ...repo,
+                        tags,
+                    };
+                })
+            )
+        }
+    }
+    
     return repos;
 }
 
+const getTagsForRepo = async <T>(owner: string, repo: string, api: Api<T>) => {
+    logger.debug(`Fetching tags for repo ${owner}/${repo}...`);
+    const { durationMs, data: tags } = await measure(() =>
+        paginate((page) => api.repos.repoListTags(owner, repo, {
+            page
+        }))
+    );
+    logger.debug(`Found ${tags.length} tags in repo ${owner}/${repo} in ${durationMs}ms.`);
+    return tags;
+}
+
+const getBranchesForRepo = async <T>(owner: string, repo: string, api: Api<T>) => {
+    logger.debug(`Fetching branches for repo ${owner}/${repo}...`);
+    const { durationMs, data: branches } = await measure(() => 
+        paginate((page) => api.repos.repoListBranches(owner, repo, {
+            page
+        }))
+    );
+    logger.debug(`Found ${branches.length} branches in repo ${owner}/${repo} in ${durationMs}ms.`);
+    return branches;
+}
+
 const getReposOwnedByUsers = async <T>(users: string[], api: Api<T>) => {
     const repos = (await Promise.all(users.map(async (user) => {
         logger.debug(`Fetching repos for user ${user}...`);

packages/backend/src/github.ts

@@ -3,12 +3,14 @@ import { GitHubConfig } from "./schemas/v2.js";
 import { createLogger } from "./logger.js";
 import { AppContext, GitRepository } from "./types.js";
 import path from 'path';
-import { excludeArchivedRepos, excludeForkedRepos, excludeReposByName, getTokenFromConfig, marshalBool } from "./utils.js";
+import { excludeArchivedRepos, excludeForkedRepos, excludeReposByName, getTokenFromConfig, marshalBool, measure } from "./utils.js";
+import micromatch from "micromatch";
 
 const logger = createLogger("GitHub");
 
 type OctokitRepository = {
     name: string,
+    id: number,
     full_name: string,
     fork: boolean,
     private: boolean,
@@ -88,7 +90,9 @@ export const getGitHubReposFromConfig = async (config: GitHubConfig, signal: Abo
                     'zoekt.archived': marshalBool(repo.archived),
                     'zoekt.fork': marshalBool(repo.fork),
                     'zoekt.public': marshalBool(repo.private === false)
-                }
+                },
+                branches: [],
+                tags: [],
             } satisfies GitRepository;
         });
 
@@ -108,9 +112,74 @@ export const getGitHubReposFromConfig = async (config: GitHubConfig, signal: Abo
 
     logger.debug(`Found ${repos.length} total repositories.`);
 
+    if (config.revisions) {
+        if (config.revisions.branches) {
+            const branchGlobs = config.revisions.branches;
+            repos = await Promise.all(
+                repos.map(async (repo) => {
+                    const [owner, name] = repo.name.split('/');
+                    let branches = (await getBranchesForRepo(owner, name, octokit, signal)).map(branch => branch.name);
+                    branches = micromatch.match(branches, branchGlobs);
+
+                    return {
+                        ...repo,
+                        branches,
+                    };
+                })
+            )
+        }
+
+        if (config.revisions.tags) {
+            const tagGlobs = config.revisions.tags;
+            repos = await Promise.all(
+                repos.map(async (repo) => {
+                    const [owner, name] = repo.name.split('/');
+                    let tags = (await getTagsForRepo(owner, name, octokit, signal)).map(tag => tag.name);
+                    tags = micromatch.match(tags, tagGlobs);
+
+                    return {
+                        ...repo,
+                        tags,
+                    };
+                })
+            )
+        }
+    }
+
     return repos;
 }
 
+const getTagsForRepo = async (owner: string, repo: string, octokit: Octokit, signal: AbortSignal) => {
+    logger.debug(`Fetching tags for repo ${owner}/${repo}...`);
+
+    const { durationMs, data: tags } = await measure(() => octokit.paginate(octokit.repos.listTags, {
+        owner,
+        repo,
+        per_page: 100,
+        request: {
+            signal
+        }
+    }));
+
+    logger.debug(`Found ${tags.length} tags for repo ${owner}/${repo} in ${durationMs}ms`);
+    return tags;
+}
+
+const getBranchesForRepo = async (owner: string, repo: string, octokit: Octokit, signal: AbortSignal) => {
+    logger.debug(`Fetching branches for repo ${owner}/${repo}...`);
+    const { durationMs, data: branches } = await measure(() => octokit.paginate(octokit.repos.listBranches, {
+        owner,
+        repo,
+        per_page: 100,
+        request: {
+            signal
+        }
+    }));
+    logger.debug(`Found ${branches.length} branches for repo ${owner}/${repo} in ${durationMs}ms`);
+    return branches;
+}
+
+
 const getReposOwnedByUsers = async (users: string[], isAuthenticated: boolean, octokit: Octokit, signal: AbortSignal) => {
     // @todo : error handling
     const repos = (await Promise.all(users.map(async (user) => {
@@ -149,7 +218,6 @@ const getReposOwnedByUsers = async (users: string[], isAuthenticated: boolean, o
 }
 
 const getReposForOrgs = async (orgs: string[], octokit: Octokit, signal: AbortSignal) => {
-    // @todo : error handling
     const repos = (await Promise.all(orgs.map(async (org) => {
         logger.debug(`Fetching repository info for org ${org}...`);
         const start = Date.now();
@@ -172,7 +240,6 @@ const getReposForOrgs = async (orgs: string[], octokit: Octokit, signal: AbortSi
 }
 
 const getRepos = async (repoList: string[], octokit: Octokit, signal: AbortSignal) => {
-    // @todo : error handling
     const repos = await Promise.all(repoList.map(async (repo) => {
         logger.debug(`Fetching repository info for ${repo}...`);
         const start = Date.now();

packages/backend/src/gitlab.ts

@@ -4,6 +4,7 @@ import { excludeArchivedRepos, excludeForkedRepos, excludeReposByName, getTokenF
 import { createLogger } from "./logger.js";
 import { AppContext, GitRepository } from "./types.js";
 import path from 'path';
+import micromatch from "micromatch";
 
 const logger = createLogger("GitLab");
 
@@ -90,7 +91,9 @@ export const getGitLabReposFromConfig = async (config: GitLabConfig, ctx: AppCon
                     'zoekt.archived': marshalBool(project.archived),
                     'zoekt.fork': marshalBool(isFork),
                     'zoekt.public': marshalBool(project.visibility === 'public'),
-                }
+                },
+                branches: [],
+                tags: [],
             } satisfies GitRepository;
         });
 
@@ -110,5 +113,41 @@ export const getGitLabReposFromConfig = async (config: GitLabConfig, ctx: AppCon
 
     logger.debug(`Found ${repos.length} total repositories.`);
 
+    if (config.revisions) {
+        if (config.revisions.branches) {
+            const branchGlobs = config.revisions.branches;
+            repos = await Promise.all(repos.map(async (repo) => {
+                logger.debug(`Fetching branches for repo ${repo.name}...`);
+                let { durationMs, data } = await measure(() => api.Branches.all(repo.name));
+                logger.debug(`Found ${data.length} branches in repo ${repo.name} in ${durationMs}ms.`);
+
+                let branches = data.map((branch) => branch.name);
+                branches = micromatch.match(branches, branchGlobs);
+
+                return {
+                    ...repo,
+                    branches,
+                };
+            }));
+        }
+
+        if (config.revisions.tags) {
+            const tagGlobs = config.revisions.tags;
+            repos = await Promise.all(repos.map(async (repo) => {
+                logger.debug(`Fetching tags for repo ${repo.name}...`);
+                let { durationMs, data } = await measure(() => api.Tags.all(repo.name));
+                logger.debug(`Found ${data.length} tags in repo ${repo.name} in ${durationMs}ms.`);
+
+                let tags = data.map((tag) => tag.name);
+                tags = micromatch.match(tags, tagGlobs);
+
+                return {
+                    ...repo,
+                    tags,
+                };
+            }));
+        }
+    }
+
     return repos;
 }

packages/backend/src/index.ts

@@ -10,7 +10,7 @@ import { AppContext, LocalRepository, GitRepository, Repository } from "./types.
 import { cloneRepository, fetchRepository } from "./git.js";
 import { createLogger } from "./logger.js";
 import { createRepository, Database, loadDB, updateRepository } from './db.js';
-import { isRemotePath, measure } from "./utils.js";
+import { arraysEqualShallow, isRemotePath, measure } from "./utils.js";
 import { REINDEX_INTERVAL_MS, RESYNC_CONFIG_INTERVAL_MS } from "./constants.js";
 import stripJsonComments from 'strip-json-comments';
 import { indexGitRepository, indexLocalRepository } from "./zoekt.js";
@@ -30,16 +30,21 @@ type Arguments = {
 const syncGitRepository = async (repo: GitRepository, ctx: AppContext) => {
     if (existsSync(repo.path)) {
         logger.info(`Fetching ${repo.id}...`);
+
         const { durationMs } = await measure(() => fetchRepository(repo, ({ method, stage , progress}) => {
             logger.info(`git.${method} ${stage} stage ${progress}% complete for ${repo.id}`)
         }));
+
         process.stdout.write('\n');
         logger.info(`Fetched ${repo.id} in ${durationMs / 1000}s`);
+
     } else {
         logger.info(`Cloning ${repo.id}...`);
+
         const { durationMs } = await measure(() => cloneRepository(repo, ({ method, stage, progress }) => {
             logger.info(`git.${method} ${stage} stage ${progress}% complete for ${repo.id}`)
         }));
+
         process.stdout.write('\n');
         logger.info(`Cloned ${repo.id} in ${durationMs / 1000}s`);
     }
@@ -55,6 +60,39 @@ const syncLocalRepository = async (repo: LocalRepository, ctx: AppContext, signa
     logger.info(`Indexed ${repo.id} in ${durationMs / 1000}s`);
 }
 
+export const isRepoReindxingRequired = (previous: Repository, current: Repository) => {
+
+    /**
+     * Checks if the any of the `revisions` properties have changed.
+     */
+    const isRevisionsChanged = () => {
+        if (previous.vcs !== 'git' || current.vcs !== 'git') {
+            return false;
+        }
+
+        return (
+            !arraysEqualShallow(previous.branches, current.branches) ||
+            !arraysEqualShallow(previous.tags, current.tags)
+        );
+    }
+
+    /**
+     * Check if the `exclude.paths` property has changed.
+     */
+    const isExcludePathsChanged = () => {
+        if (previous.vcs !== 'local' || current.vcs !== 'local') {
+            return false;
+        }
+
+        return !arraysEqualShallow(previous.excludedPaths, current.excludedPaths);
+    }
+
+    return (
+        isRevisionsChanged() ||
+        isExcludePathsChanged()
+    )
+}
+
 const syncConfig = async (configPath: string, db: Database, signal: AbortSignal, ctx: AppContext) => {
     const configContent = await (async () => {
         if (isRemotePath(configPath)) {
@@ -121,7 +159,17 @@ const syncConfig = async (configPath: string, db: Database, signal: AbortSignal,
     // Merge the repositories into the database
     for (const newRepo of configRepos) {
         if (newRepo.id in db.data.repos) {
-            await updateRepository(newRepo.id, newRepo, db);
+            const existingRepo = db.data.repos[newRepo.id];
+            const isReindexingRequired = isRepoReindxingRequired(existingRepo, newRepo);
+            if (isReindexingRequired) {
+                logger.info(`Marking ${newRepo.id} for reindexing due to configuration change.`);
+            }
+            await updateRepository(existingRepo.id, {
+                ...newRepo,
+                ...(isReindexingRequired ? {
+                    lastIndexedDate: undefined,
+                }: {})
+            }, db);
         } else {
             await createRepository(newRepo, db);
         }

packages/backend/src/schemas/v2.ts

@@ -58,6 +58,20 @@ export interface GitHubConfig {
      */
     repos?: string[];
   };
+  revisions?: GitRevisions;
+}
+/**
+ * The revisions (branches, tags) that should be included when indexing. The default branch (HEAD) is always indexed.
+ */
+export interface GitRevisions {
+  /**
+   * List of branches to include when indexing. For a given repo, only the branches that exist on the repo's remote *and* match at least one of the provided `branches` will be indexed. The default branch (HEAD) is always indexed. Glob patterns are supported.
+   */
+  branches?: string[];
+  /**
+   * List of tags to include when indexing. For a given repo, only the tags that exist on the repo's remote *and* match at least one of the provided `tags` will be indexed. Glob patterns are supported.
+   */
+  tags?: string[];
 }
 export interface GitLabConfig {
   /**
@@ -105,6 +119,7 @@ export interface GitLabConfig {
      */
     projects?: string[];
   };
+  revisions?: GitRevisions;
 }
 export interface GiteaConfig {
   /**
@@ -152,6 +167,7 @@ export interface GiteaConfig {
      */
     repos?: string[];
   };
+  revisions?: GitRevisions;
 }
 export interface LocalConfig {
   /**

packages/backend/src/types.ts

@@ -13,6 +13,8 @@ interface BaseRepository {
 export interface GitRepository extends BaseRepository {
     vcs: 'git';
     cloneUrl: string;
+    branches: string[];
+    tags: string[];
     gitConfigMetadata?: Record<string, string>;
 }
 

packages/backend/src/utils.ts

@@ -74,3 +74,11 @@ export const resolvePathRelativeToConfig = (localPath: string, configPath: strin
 
     return absolutePath;
 }
+
+export const arraysEqualShallow = <T>(a?: T[], b?: T[]) => {
+    if (a === b) return true;
+    if (a === undefined || b === undefined) return false;
+    if (a.length !== b.length) return false;
+
+    return a.every(item => b.includes(item)) && b.every(item => a.includes(item));
+}

packages/backend/src/zoekt.ts

@@ -4,8 +4,16 @@ import { AppContext, GitRepository, LocalRepository } from "./types.js";
 const ALWAYS_EXCLUDED_DIRS = ['.git', '.hg', '.svn'];
 
 export const indexGitRepository = async (repo: GitRepository, ctx: AppContext) => {
+    const revisions = [
+        'HEAD',
+        ...repo.branches ?? [],
+        ...repo.tags ?? [],
+    ];
+
+    const command = `zoekt-git-index -index ${ctx.indexPath} -branches ${revisions.join(',')} ${repo.path}`;
+
     return new Promise<{ stdout: string, stderr: string }>((resolve, reject) => {
-        exec(`zoekt-git-index -index ${ctx.indexPath} ${repo.path}`, (error, stdout, stderr) => {
+        exec(command, (error, stdout, stderr) => {
             if (error) {
                 reject(error);
                 return;

packages/web/package.json

@@ -61,6 +61,7 @@
     "server-only": "^0.0.1",
     "sharp": "^0.33.5",
     "tailwind-merge": "^2.5.2",
+    "tailwind-scrollbar-hide": "^1.1.7",
     "tailwindcss-animate": "^1.0.7",
     "usehooks-ts": "^3.1.0",
     "zod": "^3.23.8"

packages/web/src/app/api/(client)/client.ts

@@ -1,5 +1,5 @@
 import { fileSourceResponseSchema, listRepositoriesResponseSchema, searchResponseSchema } from "@/lib/schemas";
-import { FileSourceResponse, ListRepositoriesResponse, SearchRequest, SearchResponse } from "@/lib/types";
+import { FileSourceRequest, FileSourceResponse, ListRepositoriesResponse, SearchRequest, SearchResponse } from "@/lib/types";
 
 export const search = async (body: SearchRequest): Promise<SearchResponse> => {
     const result = await fetch(`/api/search`, {
@@ -13,16 +13,13 @@ export const search = async (body: SearchRequest): Promise<SearchResponse> => {
     return searchResponseSchema.parse(result);
 }
 
-export const fetchFileSource = async (fileName: string, repository: string): Promise<FileSourceResponse> => {
+export const fetchFileSource = async (body: FileSourceRequest): Promise<FileSourceResponse> => {
     const result = await fetch(`/api/source`, {
         method: "POST",
         headers: {
             "Content-Type": "application/json",
         },
-        body: JSON.stringify({
+        body: JSON.stringify(body),
-            fileName,
-            repository,
-        }),
     }).then(response => response.json());
 
     return fileSourceResponseSchema.parse(result);

packages/web/src/app/page.tsx

@@ -76,7 +76,7 @@ export default async function Home() {
                                 <Query query="lang:typescript"><Highlight>lang:</Highlight>typescript</Query> <QueryExplanation>(by language)</QueryExplanation>
                             </QueryExample>
                             <QueryExample>
-                                <Query query="branch:HEAD"><Highlight>branch:</Highlight>HEAD</Query> <QueryExplanation>(by branch)</QueryExplanation>
+                                <Query query="revision:HEAD"><Highlight>revision:</Highlight>HEAD</Query> <QueryExplanation>(by branch or tag)</QueryExplanation>
                             </QueryExample>
                         </HowToSection>
                         <HowToSection

packages/web/src/app/repos/columns.tsx

@@ -55,7 +55,7 @@ export const columns: ColumnDef<RepositoryColumnInfo>[] = [
             }
 
             return (
-                <div className="flex flex-col gap-2">
+                <div className="flex flex-col gap-2 max-h-32 overflow-scroll scrollbar-hide">
                     {branches.map(({ name, version }, index) => {
                         const shortVersion = version.substring(0, 8);
                         return (

packages/web/src/app/search/components/codePreviewPanel/codePreview.tsx

@@ -124,7 +124,7 @@ export const CodePreview = ({
                 {/* File path */}
                 <div className="flex-1 overflow-hidden">
                     <span
-                        className={clsx("block truncate-start", {
+                        className={clsx("block truncate-start text-sm font-mono", {
                             "cursor-pointer text-blue-500 hover:underline": file?.link
                         })}
                         onClick={() => {

packages/web/src/app/search/components/codePreviewPanel/index.tsx

@@ -21,16 +21,24 @@ export const CodePreviewPanel = ({
 }: CodePreviewPanelProps) => {
 
     const { data: file } = useQuery({
-        queryKey: ["source", fileMatch?.FileName, fileMatch?.Repository],
+        queryKey: ["source", fileMatch?.FileName, fileMatch?.Repository, fileMatch?.Branches],
         queryFn: async (): Promise<CodePreviewFile | undefined> => {
             if (!fileMatch) {
                 return undefined;
             }
 
-            return fetchFileSource(fileMatch.FileName, fileMatch.Repository)
+            // If there are multiple branches pointing to the same revision of this file, it doesn't
+            // matter which branch we use here, so use the first one.
+            const branch = fileMatch.Branches && fileMatch.Branches.length > 0 ? fileMatch.Branches[0] : undefined;
+
+            return fetchFileSource({
+                fileName: fileMatch.FileName,
+                repository: fileMatch.Repository,
+                branch,
+            })
                 .then(({ source }) => {
                     // @todo : refector this to use the templates provided by zoekt.
-                    const link = getCodeHostFilePreviewLink(fileMatch.Repository, fileMatch.FileName)
+                    const link = getCodeHostFilePreviewLink(fileMatch.Repository, fileMatch.FileName, branch);
 
                     const decodedSource = base64Decode(source);
 

packages/web/src/app/search/components/searchResultsPanel/fileMatchContainer.tsx

@@ -17,6 +17,7 @@ interface FileMatchContainerProps {
     onMatchIndexChanged: (matchIndex: number) => void;
     showAllMatches: boolean;
     onShowAllMatchesButtonClicked: () => void;
+    isBranchFilteringEnabled: boolean;
 }
 
 export const FileMatchContainer = ({
@@ -25,6 +26,7 @@ export const FileMatchContainer = ({
     onMatchIndexChanged,
     showAllMatches,
     onShowAllMatchesButtonClicked,
+    isBranchFilteringEnabled,
 }: FileMatchContainerProps) => {
 
     const matchCount = useMemo(() => {
@@ -90,6 +92,14 @@ export const FileMatchContainer = ({
         onMatchIndexChanged(matchIndex);
     }, [matches, onMatchIndexChanged, onOpenFile]);
 
+    const branches = useMemo(() => {
+        if (!file.Branches) {
+            return [];
+        }
+
+        return file.Branches;
+    }, [file.Branches]);
+
 
     return (
         <div>
@@ -114,9 +124,18 @@ export const FileMatchContainer = ({
                     >
                         {repoName}
                     </span>
+                    {isBranchFilteringEnabled && branches.length > 0 && (
+                        <span
+                            className="text-xs font-semibold text-gray-500 dark:text-gray-400 mt-0.5"
+                            title={branches.join(", ")}
+                        >
+                            {`@ ${branches[0]}`}
+                            {branches.length > 1 && ` (+ ${branches.length - 1})`}
+                        </span>
+                    )}
                     <span>·</span>
                     <div className="flex-1 flex items-center overflow-hidden">
-                        <span className="inline-block w-full truncate-start">
+                        <span className="inline-block w-full truncate-start font-mono text-sm">
                             {!fileNameRange ?
                                 file.FileName
                             : (

packages/web/src/app/search/components/searchResultsPanel/index.tsx

@@ -11,6 +11,7 @@ interface SearchResultsPanelProps {
     onMatchIndexChanged: (matchIndex: number) => void;
     isLoadMoreButtonVisible: boolean;
     onLoadMoreButtonClicked: () => void;
+    isBranchFilteringEnabled: boolean;
 }
 
 const ESTIMATED_LINE_HEIGHT_PX = 20;
@@ -23,6 +24,7 @@ export const SearchResultsPanel = ({
     onMatchIndexChanged,
     isLoadMoreButtonVisible,
     onLoadMoreButtonClicked,
+    isBranchFilteringEnabled,
 }: SearchResultsPanelProps) => {
     const parentRef = useRef<HTMLDivElement>(null);
     const [showAllMatchesStates, setShowAllMatchesStates] = useState(Array(fileMatches.length).fill(false));
@@ -145,6 +147,7 @@ export const SearchResultsPanel = ({
                             onShowAllMatchesButtonClicked={() => {
                                 onShowAllMatchesButtonClicked(virtualRow.index);
                             }}
+                            isBranchFilteringEnabled={isBranchFilteringEnabled}
                         />
                     </div>
                 ))}

packages/web/src/app/search/page.tsx

@@ -77,35 +77,55 @@ export default function SearchPage() {
         });
     }, [captureEvent, searchResponse]);
 
-    const { fileMatches, searchDurationMs } = useMemo((): { fileMatches: SearchResultFile[], searchDurationMs: number } => {
+    const { fileMatches, searchDurationMs, totalMatchCount, isBranchFilteringEnabled } = useMemo(() => {
         if (!searchResponse) {
             return {
                 fileMatches: [],
                 searchDurationMs: 0,
+                totalMatchCount: 0,
+                isBranchFilteringEnabled: false,
             };
         }
 
-        return {
+        const isBranchFilteringEnabled = searchResponse.isBranchFilteringEnabled;
-            fileMatches: searchResponse.Result.Files ?? [],
+        let fileMatches = searchResponse.Result.Files ?? [];
-            searchDurationMs: Math.round(searchResponse.Result.Duration / 1000000),
+
+        // We only want to show matches for the default branch when
+        // the user isn't explicitly filtering by branch.
+        if (!isBranchFilteringEnabled) {
+            fileMatches = fileMatches.filter(match => {
+                // @note : this case handles local repos that don't have any branches.
+                if (!match.Branches) {
+                    return true;
+                }
+
+                return match.Branches.includes("HEAD");
+            });
         }
-    }, [searchResponse]);
+
+        return {
+            fileMatches,
+            searchDurationMs: Math.round(searchResponse.Result.Duration / 1000000),
+            totalMatchCount: searchResponse.Result.MatchCount,
+            isBranchFilteringEnabled,
+        }
+    }, [searchResponse, searchQuery]);
 
     const isMoreResultsButtonVisible = useMemo(() => {
-        return searchResponse && searchResponse.Result.MatchCount > maxMatchDisplayCount;
+        return totalMatchCount > maxMatchDisplayCount;
-    }, [searchResponse, maxMatchDisplayCount]);
+    }, [totalMatchCount, maxMatchDisplayCount]);
 
     const numMatches = useMemo(() => {
         // Accumualtes the number of matches across all files
-        return searchResponse?.Result.Files?.reduce(
+        return fileMatches.reduce(
             (acc, file) =>
                 acc + file.ChunkMatches.reduce(
                     (acc, chunk) => acc + chunk.Ranges.length,
                     0,
                 ),
             0,
-        ) ?? 0;
+        );
-    }, [searchResponse]);
+    }, [fileMatches]);
 
     const onLoadMoreResults = useCallback(() => {
         const url = createPathWithQueryParams('/search',
@@ -151,8 +171,8 @@ export default function SearchPage() {
                 {!isLoading && (
                     <div className="bg-accent py-1 px-2 flex flex-row items-center gap-4">
                         {
-                            fileMatches.length > 0 && searchResponse ? (
+                            fileMatches.length > 0 ? (
-                                <p className="text-sm font-medium">{`[${searchDurationMs} ms] Displaying ${numMatches} of ${searchResponse.Result.MatchCount} matches in ${fileMatches.length} ${fileMatches.length > 1 ? 'files' : 'file'}`}</p>
+                                <p className="text-sm font-medium">{`[${searchDurationMs} ms] Found ${numMatches} matches in ${fileMatches.length} ${fileMatches.length > 1 ? 'files' : 'file'}`}</p>
                             ) : (
                                 <p className="text-sm font-medium">No results</p>
                             )
@@ -180,6 +200,7 @@ export default function SearchPage() {
                     fileMatches={fileMatches}
                     isMoreResultsButtonVisible={isMoreResultsButtonVisible}
                     onLoadMoreResults={onLoadMoreResults}
+                    isBranchFilteringEnabled={isBranchFilteringEnabled}
                 />
             )}
         </div>
@@ -190,12 +211,14 @@ interface PanelGroupProps {
     fileMatches: SearchResultFile[];
     isMoreResultsButtonVisible?: boolean;
     onLoadMoreResults: () => void;
+    isBranchFilteringEnabled: boolean;
 }
 
 const PanelGroup = ({
     fileMatches,
     isMoreResultsButtonVisible,
     onLoadMoreResults,
+    isBranchFilteringEnabled,
 }: PanelGroupProps) => {
     const [selectedMatchIndex, setSelectedMatchIndex] = useState(0);
     const [selectedFile, setSelectedFile] = useState<SearchResultFile | undefined>(undefined);
@@ -253,6 +276,7 @@ const PanelGroup = ({
                         }}
                         isLoadMoreButtonVisible={!!isMoreResultsButtonVisible}
                         onLoadMoreButtonClicked={onLoadMoreResults}
+                        isBranchFilteringEnabled={isBranchFilteringEnabled}
                     />
                 ) : (
                     <div className="flex flex-col items-center justify-center h-full">

packages/web/src/lib/schemas.ts

@@ -47,7 +47,7 @@ export const searchResponseStats = {
 }
 
 // @see : https://github.com/sourcebot-dev/zoekt/blob/3780e68cdb537d5a7ed2c84d9b3784f80c7c5d04/api.go#L497
-export const searchResponseSchema = z.object({
+export const zoektSearchResponseSchema = z.object({
     Result: z.object({
         ...searchResponseStats,
         Files: z.array(z.object({
@@ -71,9 +71,16 @@ export const searchResponseSchema = z.object({
     }),
 });
 
+export const searchResponseSchema = z.object({
+    ...zoektSearchResponseSchema.shape,
+    // Flag when a branch filter was used (e.g., `branch:`, `revision:`, etc.).
+    isBranchFilteringEnabled: z.boolean(),
+});
+
 export const fileSourceRequestSchema = z.object({
     fileName: z.string(),
-    repository: z.string()
+    repository: z.string(),
+    branch: z.string().optional(),
 });
 
 export const fileSourceResponseSchema = z.object({

packages/web/src/lib/server/searchService.ts

@@ -1,12 +1,45 @@
 import escapeStringRegexp from "escape-string-regexp";
 import { SHARD_MAX_MATCH_COUNT, TOTAL_MAX_MATCH_COUNT } from "../environment";
-import { listRepositoriesResponseSchema, searchResponseSchema } from "../schemas";
+import { listRepositoriesResponseSchema, searchResponseSchema, zoektSearchResponseSchema } from "../schemas";
 import { FileSourceRequest, FileSourceResponse, ListRepositoriesResponse, SearchRequest, SearchResponse } from "../types";
 import { fileNotFound, invalidZoektResponse, ServiceError, unexpectedError } from "../serviceError";
 import { isServiceError } from "../utils";
 import { zoektFetch } from "./zoektClient";
 
+// List of supported query prefixes in zoekt.
+// @see : https://github.com/sourcebot-dev/zoekt/blob/main/query/parse.go#L417
+enum zoektPrefixes {
+    archived = "archived:",
+    branchShort = "b:",
+    branch =  "branch:",
+    caseShort =  "c:",
+    case =  "case:",
+    content =  "content:",
+    fileShort =  "f:",
+    file =  "file:",
+    fork =  "fork:",
+    public =  "public:",
+    repoShort =  "r:",
+    repo =  "repo:",
+    regex =  "regex:",
+    lang =  "lang:",
+    sym =  "sym:",
+    typeShort =  "t:",
+    type =  "type:",
+}
+
+// Mapping of additional "alias" prefixes to zoekt prefixes.
+const aliasPrefixMappings: Record<string, zoektPrefixes> = {
+    "rev:": zoektPrefixes.branch,
+    "revision:": zoektPrefixes.branch,
+}
+
 export const search = async ({ query, maxMatchDisplayCount, whole }: SearchRequest): Promise<SearchResponse | ServiceError> => {
+    // Replace any alias prefixes with their corresponding zoekt prefixes.
+    for (const [prefix, zoektPrefix] of Object.entries(aliasPrefixMappings)) {
+        query = query.replaceAll(prefix, zoektPrefix);
+    }
+
     const body = JSON.stringify({
         q: query,
         // @see: https://github.com/sourcebot-dev/zoekt/blob/main/api.go#L892
@@ -31,21 +64,34 @@ export const search = async ({ query, maxMatchDisplayCount, whole }: SearchReque
     }
 
     const searchBody = await searchResponse.json();
-    const parsedSearchResponse = searchResponseSchema.safeParse(searchBody);
+    const parsedSearchResponse = zoektSearchResponseSchema.safeParse(searchBody);
     if (!parsedSearchResponse.success) {
         console.error(`Failed to parse zoekt response. Error: ${parsedSearchResponse.error}`);
         return unexpectedError(`Something went wrong while parsing the response from zoekt`);
     }
 
-    return parsedSearchResponse.data;
+    const isBranchFilteringEnabled = (
+        query.includes(zoektPrefixes.branch) ||
+        query.includes(zoektPrefixes.branchShort)
+    )
+
+    return {
+        ...parsedSearchResponse.data,
+        isBranchFilteringEnabled,
+    }
 }
 
-export const getFileSource = async ({ fileName, repository }: FileSourceRequest): Promise<FileSourceResponse | ServiceError> => {
+export const getFileSource = async ({ fileName, repository, branch }: FileSourceRequest): Promise<FileSourceResponse | ServiceError> => {
     const escapedFileName = escapeStringRegexp(fileName);
     const escapedRepository = escapeStringRegexp(repository);
     
+    let query = `file:${escapedFileName} repo:^${escapedRepository}$`;
+    if (branch) {
+        query = query.concat(` branch:${branch}`);
+    }
+
     const searchResponse = await search({
-        query: `file:${escapedFileName} repo:^${escapedRepository}$`,
+        query,
         maxMatchDisplayCount: 1,
         whole: true,
     });

packages/web/src/lib/types.ts

@@ -3,7 +3,9 @@ import { fileSourceRequestSchema, fileSourceResponseSchema, listRepositoriesResp
 
 export type KeymapType = "default" | "vim";
 
+export type SearchRequest = z.infer<typeof searchRequestSchema>;
 export type SearchResponse = z.infer<typeof searchResponseSchema>;
+
 export type SearchResult = SearchResponse["Result"];
 export type SearchResultFile = NonNullable<SearchResult["Files"]>[number];
 export type SearchResultFileMatch = SearchResultFile["ChunkMatches"][number];
@@ -15,7 +17,6 @@ export type FileSourceResponse = z.infer<typeof fileSourceResponseSchema>;
 
 export type ListRepositoriesResponse = z.infer<typeof listRepositoriesResponseSchema>;
 export type Repository = z.infer<typeof repositorySchema>;
-export type SearchRequest = z.infer<typeof searchRequestSchema>;
 
 export enum SearchQueryParams {
     query = "query",

packages/web/src/lib/utils.ts

@@ -73,19 +73,19 @@ export const getRepoCodeHostInfo = (repoName: string): CodeHostInfo | undefined
     return undefined;
 }
 
-export const getCodeHostFilePreviewLink = (repoName: string, filePath: string): string | undefined => {
+export const getCodeHostFilePreviewLink = (repoName: string, filePath: string, branch: string = "HEAD"): string | undefined => {
     const info = getRepoCodeHostInfo(repoName);
 
     if (info?.type === "github") {
-        return `${info.repoLink}/blob/HEAD/${filePath}`;
+        return `${info.repoLink}/blob/${branch}/${filePath}`;
     }
 
     if (info?.type === "gitlab") {
-        return `${info.repoLink}/-/blob/HEAD/${filePath}`;
+        return `${info.repoLink}/-/blob/${branch}/${filePath}`;
     }
 
     if (info?.type === "gitea") {
-        return `${info.repoLink}/src/branch/HEAD/${filePath}`;
+        return `${info.repoLink}/src/branch/${branch}/${filePath}`;
     }
 
     return undefined;

packages/web/tailwind.config.ts

@@ -74,7 +74,10 @@ const config = {
       },
     },
   },
-  plugins: [require("tailwindcss-animate")],
+  plugins: [
+    require("tailwindcss-animate"),
+    require('tailwind-scrollbar-hide')
+  ],
 } satisfies Config
 
 export default config

schemas/v2/index.json

@@ -24,6 +24,47 @@
                 }
             ]
         },
+        "GitRevisions": {
+            "type": "object",
+            "description": "The revisions (branches, tags) that should be included when indexing. The default branch (HEAD) is always indexed.",
+            "properties": {
+                "branches": {
+                    "type": "array",
+                    "description": "List of branches to include when indexing. For a given repo, only the branches that exist on the repo's remote *and* match at least one of the provided `branches` will be indexed. The default branch (HEAD) is always indexed. Glob patterns are supported.",
+                    "items": {
+                        "type": "string"
+                    },
+                    "examples": [
+                        [
+                            "main",
+                            "release/*"
+                        ],
+                        [
+                            "**"
+                        ]
+                    ],
+                    "default": []
+                },
+                "tags": {
+                    "type": "array",
+                    "description": "List of tags to include when indexing. For a given repo, only the tags that exist on the repo's remote *and* match at least one of the provided `tags` will be indexed. Glob patterns are supported.",
+                    "items": {
+                        "type": "string"
+                    },
+                    "examples": [
+                        [
+                            "latest",
+                            "v2.*.*"
+                        ],
+                        [
+                            "**"
+                        ]
+                    ],
+                    "default": []
+                }
+            },
+            "additionalProperties": false
+        },
         "GitHubConfig": {
             "type": "object",
             "properties": {
@@ -113,6 +154,9 @@
                         }
                     },
                     "additionalProperties": false
+                },
+                "revisions": {
+                    "$ref": "#/definitions/GitRevisions"
                 }
             },
             "required": [
@@ -207,6 +251,9 @@
                         }
                     },
                     "additionalProperties": false
+                },
+                "revisions": {
+                    "$ref": "#/definitions/GitRevisions"
                 }
             },
             "required": [
@@ -297,6 +344,9 @@
                         }
                     },
                     "additionalProperties": false
+                },
+                "revisions": {
+                    "$ref": "#/definitions/GitRevisions"
                 }
             },
             "required": [

yarn.lock

@@ -1327,6 +1327,11 @@
   resolved "https://registry.yarnpkg.com/@types/argparse/-/argparse-2.0.16.tgz#3bb7ccd2844b3a8bcd6efbd217f6c0ea06a80d22"
   integrity sha512-aMqBra2JlqpFeCWOinCtpRpiCkPIXH8hahW2+FkGzvWjfE5sAqtOcrjN5DRcMnTQqFDe6gb1CVYuGnBH0lhXwA==
 
+"@types/braces@*":
+  version "3.0.4"
+  resolved "https://registry.yarnpkg.com/@types/braces/-/braces-3.0.4.tgz#403488dc1c8d0db288270d3bbf0ce5f9c45678b4"
+  integrity sha512-0WR3b8eaISjEW7RpZnclONaLFDf7buaowRHdqLp4vLj54AsSAYWfh3DRbfiYJY9XDxMgx1B4sE1Afw2PGpuHOA==
+
 "@types/json-schema@^7.0.15":
   version "7.0.15"
   resolved "https://registry.yarnpkg.com/@types/json-schema/-/json-schema-7.0.15.tgz#596a1747233694d50f6ad8a7869fcb6f56cf5841"
@@ -1342,6 +1347,13 @@
   resolved "https://registry.yarnpkg.com/@types/lodash/-/lodash-4.17.10.tgz#64f3edf656af2fe59e7278b73d3e62404144a6e6"
   integrity sha512-YpS0zzoduEhuOWjAotS6A5AVCva7X4lVlYLF0FYHAY9sdraBfnatttHItlWeZdGhuEkf+OzMNg2ZYAx8t+52uQ==
 
+"@types/micromatch@^4.0.9":
+  version "4.0.9"
+  resolved "https://registry.yarnpkg.com/@types/micromatch/-/micromatch-4.0.9.tgz#8e5763a8c1fc7fbf26144d9215a01ab0ff702dbb"
+  integrity sha512-7V+8ncr22h4UoYRLnLXSpTxjQrNUXtWHGeMPRJt1nULXI57G9bIcpyrHlmrQ7QK24EyyuXvYcSSWAM8GA9nqCg==
+  dependencies:
+    "@types/braces" "*"
+
 "@types/node@^20":
   version "20.16.10"
   resolved "https://registry.yarnpkg.com/@types/node/-/node-20.16.10.tgz#0cc3fdd3daf114a4776f54ba19726a01c907ef71"
@@ -3444,7 +3456,7 @@ merge2@^1.3.0, merge2@^1.4.1:
   resolved "https://registry.yarnpkg.com/merge2/-/merge2-1.4.1.tgz#4368892f885e907455a6fd7dc55c0c9d404990ae"
   integrity sha512-8q7VEgMJW4J8tcfVPy8g09NcQwZdbwFEqhe/WZkoIzjn/3TGDwtOCYtXGxA3O8tPzpczCCDgv+P2P5y00ZJOOg==
 
-micromatch@^4.0.4, micromatch@^4.0.5:
+micromatch@^4.0.4, micromatch@^4.0.5, micromatch@^4.0.8:
   version "4.0.8"
   resolved "https://registry.yarnpkg.com/micromatch/-/micromatch-4.0.8.tgz#d66fa18f3a47076789320b9b1af32bd86d9fa202"
   integrity sha512-PXwfBhYu0hBCPw8Dn0E+WDYb7af3dSLVWKi3HGv84IdF4TyFoC0ysxFd0Goxw7nSv4T/PzEJQxsYsEiFCKo2BA==
@@ -4544,6 +4556,11 @@ tailwind-merge@^2.5.2:
   resolved "https://registry.yarnpkg.com/tailwind-merge/-/tailwind-merge-2.5.3.tgz#579546e14ddda24462e0303acd8798c50f5511bb"
   integrity sha512-d9ZolCAIzom1nf/5p4LdD5zvjmgSxY0BGgdSvmXIoMYAiPdAW/dSpP7joCDYFY7r/HkEa2qmPtkgsu0xjQeQtw==
 
+tailwind-scrollbar-hide@^1.1.7:
+  version "1.1.7"
+  resolved "https://registry.yarnpkg.com/tailwind-scrollbar-hide/-/tailwind-scrollbar-hide-1.1.7.tgz#90b481fb2e204030e3919427416650c54f56f847"
+  integrity sha512-X324n9OtpTmOMqEgDUEA/RgLrNfBF/jwJdctaPZDzB3mppxJk7TLIDmOreEDm1Bq4R9LSPu4Epf8VSdovNU+iA==
+
 tailwindcss-animate@^1.0.7:
   version "1.0.7"
   resolved "https://registry.yarnpkg.com/tailwindcss-animate/-/tailwindcss-animate-1.0.7.tgz#318b692c4c42676cc9e67b19b78775742388bef4"