feat(web): Streamed code search

.env.development

@@ -6,8 +6,6 @@ DATABASE_URL="postgresql://postgres:postgres@localhost:5432/postgres"
 ZOEKT_WEBSERVER_URL="http://localhost:6070"
 # The command to use for generating ctags.
 CTAGS_COMMAND=ctags
-# logging, strict
-SRC_TENANT_ENFORCEMENT_MODE=strict
 
 # Auth.JS
 # You can generate a new secret with:

CHANGELOG.md

@@ -7,9 +7,23 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+<!-- Bump @sourcebot/mcp since there are breaking changes to the api in this release -->
+
 ### Added
+- Added support for streaming code search results. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+- Added buttons to toggle case sensitivity and regex patterns. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
 - Added counts to members, requets, and invites tabs in the members settings. [#621](https://github.com/sourcebot-dev/sourcebot/pull/621)
 
+### Changed
+- Changed the default search behaviour to match patterns as substrings and **not** regular expressions. Regular expressions can be used by toggling the regex button in search bar. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+- Renamed `public` query prefix to `visibility`. Allowed values for `visibility` are `public`, `private`, and `any`. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+- Changed `archived` query prefix to accept values `yes`, `no`, and `only`. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+
+### Removed
+- Removed `case` query prefix. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+- Removed `branch` and `b` query prefixes. Please use `rev:` instead. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+- Removed `regex` query prefix. [#623](https://github.com/sourcebot-dev/sourcebot/pull/623)
+
 ### Fixed
 - Fixed spurious infinite loads with explore panel, file tree, and file search command. [#617](https://github.com/sourcebot-dev/sourcebot/pull/617)
 - Wipe search context on init if entitlement no longer exists [#618](https://github.com/sourcebot-dev/sourcebot/pull/618)

Dockerfile

@@ -43,10 +43,12 @@ COPY .yarn ./.yarn
 COPY ./packages/db ./packages/db
 COPY ./packages/schemas ./packages/schemas
 COPY ./packages/shared ./packages/shared
+COPY ./packages/queryLanguage ./packages/queryLanguage
 
 RUN yarn workspace @sourcebot/db install
 RUN yarn workspace @sourcebot/schemas install
 RUN yarn workspace @sourcebot/shared install
+RUN yarn workspace @sourcebot/query-language install
 # ------------------------------------
 
 # ------ Build Web ------
@@ -92,6 +94,7 @@ COPY --from=shared-libs-builder /app/node_modules ./node_modules
 COPY --from=shared-libs-builder /app/packages/db ./packages/db
 COPY --from=shared-libs-builder /app/packages/schemas ./packages/schemas
 COPY --from=shared-libs-builder /app/packages/shared ./packages/shared
+COPY --from=shared-libs-builder /app/packages/queryLanguage ./packages/queryLanguage
 
 # Fixes arm64 timeouts
 RUN yarn workspace @sourcebot/web install
@@ -130,6 +133,7 @@ COPY --from=shared-libs-builder /app/node_modules ./node_modules
 COPY --from=shared-libs-builder /app/packages/db ./packages/db
 COPY --from=shared-libs-builder /app/packages/schemas ./packages/schemas
 COPY --from=shared-libs-builder /app/packages/shared ./packages/shared
+COPY --from=shared-libs-builder /app/packages/queryLanguage ./packages/queryLanguage
 RUN yarn workspace @sourcebot/backend install
 RUN yarn workspace @sourcebot/backend build
 
@@ -173,7 +177,6 @@ ENV DATA_DIR=/data
 ENV DATA_CACHE_DIR=$DATA_DIR/.sourcebot
 ENV DATABASE_DATA_DIR=$DATA_CACHE_DIR/db
 ENV REDIS_DATA_DIR=$DATA_CACHE_DIR/redis
-ENV SRC_TENANT_ENFORCEMENT_MODE=strict
 ENV SOURCEBOT_PUBLIC_KEY_PATH=/app/public.pem
 
 # Valid values are: debug, info, warn, error
@@ -217,6 +220,9 @@ COPY --from=zoekt-builder \
 /cmd/zoekt-index \
 /usr/local/bin/
 
+# Copy zoekt proto files (needed for gRPC client at runtime)
+COPY vendor/zoekt/grpc/protos /app/vendor/zoekt/grpc/protos
+
 # Copy all of the things
 COPY --from=web-builder /app/packages/web/public ./packages/web/public
 COPY --from=web-builder /app/packages/web/.next/standalone ./
@@ -229,6 +235,7 @@ COPY --from=shared-libs-builder /app/node_modules ./node_modules
 COPY --from=shared-libs-builder /app/packages/db ./packages/db
 COPY --from=shared-libs-builder /app/packages/schemas ./packages/schemas
 COPY --from=shared-libs-builder /app/packages/shared ./packages/shared
+COPY --from=shared-libs-builder /app/packages/queryLanguage ./packages/queryLanguage
 
 # Fixes git "dubious ownership" issues when the volume is mounted with different permissions to the container.
 RUN git config --global safe.directory "*"

docs/docs/features/search/syntax-reference.mdx

@@ -4,32 +4,51 @@ title: Writing search queries
 
 Sourcebot uses a powerful regex-based query language that enabled precise code search within large codebases.
 
-
 ## Syntax reference guide
 
-Queries consist of space-separated regular expressions. Wrapping expressions in `""` combines them. By default, a file must have at least one match for each expression to be included.
+Queries consist of space-separated search patterns that are matched against file contents. A file must have at least one match for each expression to be included. Queries can optionally contain search filters to further refine the search results.
+
+## Keyword search (default)
+
+Keyword search matches search patterns exactly in file contents. Wrapping search patterns in `""` combines them as a single expression.
 
 | Example | Explanation |
 | :--- | :--- |
-| `foo` | Match files with regex `/foo/` |
-| `foo bar` | Match files with regex `/foo/` **and** `/bar/` |
-| `"foo bar"` | Match files with regex `/foo bar/` |
+| `foo` | Match files containing the keyword `foo` |
+| `foo bar` | Match files containing both `foo` **and** `bar` |
+| `"foo bar"` | Match files containing the phrase `foo bar` |
+| `"foo \"bar\""` | Match files containing `foo "bar"` exactly (escaped quotes) |
+
+## Regex search
 
-Multiple expressions can be or'd together with `or`, negated with `-`, or grouped with `()`.
+Toggle the regex button (`.*`) in the search bar to interpret search patterns as regular expressions.
 
 | Example | Explanation |
 | :--- | :--- |
-| `foo or bar` | Match files with regex `/foo/` **or** `/bar/` |
-| `foo -bar` | Match files with regex `/foo/` but **not** `/bar/` |
-| `foo (bar or baz)` | Match files with regex `/foo/` **and** either `/bar/` **or** `/baz/` |
+| `foo` | Match files with regex `/foo/` |
+| `foo.*bar` | Match files with regex `/foo.*bar/` (foo followed by any characters, then bar) |
+| `^function\s+\w+` | Match files with regex `/^function\s+\w+/` (function at start of line, followed by whitespace and word characters) |
+| `"foo bar"` | Match files with regex `/foo bar/`. Quotes are not matched. |
 
-Expressions can be prefixed with certain keywords to modify search behavior. Some keywords can be negated using the `-` prefix.
+## Search filters
+
+Search queries (keyword or regex) can include multiple search filters to further refine the search results. Some filters can be negated using the `-` prefix.
 
 | Prefix | Description | Example |
 | :--- | :--- | :--- |
 | `file:` | Filter results from filepaths that match the regex. By default all files are searched. | `file:README` - Filter results to filepaths that match regex `/README/`<br/>`file:"my file"` - Filter results to filepaths that match regex `/my file/`<br/>`-file:test\.ts$` - Ignore results from filepaths match regex `/test\.ts$/` |
-| `repo:` | Filter results from repos that match the regex. By default all repos are searched. | `repo:linux` - Filter results to repos that match regex `/linux/`<br/>`-repo:^web/.*` - Ignore results from repos that match regex `/^web\/.*` |
+| `repo:` | Filter results from repos that match the regex. By default all repos are searched. | `repo:linux` - Filter results to repos that match regex `/linux/`<br/>`-repo:^web/.*` - Ignore results from repos that match regex `/^web\/.*/` |
 | `rev:` | Filter results from a specific branch or tag. By default **only** the default branch is searched. | `rev:beta` - Filter results to branches that match regex `/beta/` |
 | `lang:` | Filter results by language (as defined by [linguist](https://github.com/github-linguist/linguist/blob/main/lib/linguist/languages.yml)). By default all languages are searched. | `lang:TypeScript` - Filter results to TypeScript files<br/>`-lang:YAML` - Ignore results from YAML files |
 | `sym:` | Match symbol definitions created by [universal ctags](https://ctags.io/) at index time. | `sym:\bmain\b` - Filter results to symbols that match regex `/\bmain\b/` |
-| `context:` | Filter results to a predefined [search context](/docs/features/search/search-contexts). | `context:web` - Filter results to the web context<br/>`-context:pipelines` - Ignore results from the pipelines context |
+| `context:` | Filter results to a predefined [search context](/docs/features/search/search-contexts). | `context:web` - Filter results to the web context<br/>`-context:pipelines` - Ignore results from the pipelines context |
+
+## Boolean operators & grouping
+
+By default, space-separated expressions are and'd together. Using the `or` keyword as well as parentheses `()` can be used to create more complex boolean logic. Parentheses can be negated using the `-` prefix.
+
+| Example | Explanation |
+| :--- | :--- |
+| `foo or bar` | Match files containing `foo` **or** `bar` |
+| `foo (bar or baz)` | Match files containing `foo` **and** either `bar` **or** `baz`. |
+| `-(foo) bar` | Match files containing `bar` **and not** `foo`. |

package.json

@@ -18,7 +18,7 @@
         "dev:prisma:studio": "yarn with-env yarn workspace @sourcebot/db prisma:studio",
         "dev:prisma:migrate:reset": "yarn with-env yarn workspace @sourcebot/db prisma:migrate:reset",
         "dev:prisma:db:push": "yarn with-env yarn workspace @sourcebot/db prisma:db:push",
-        "build:deps": "yarn workspaces foreach --recursive --topological --from '{@sourcebot/schemas,@sourcebot/db,@sourcebot/shared}' run build"
+        "build:deps": "yarn workspaces foreach --recursive --topological --from '{@sourcebot/schemas,@sourcebot/db,@sourcebot/shared,@sourcebot/query-language}' run build"
     },
     "devDependencies": {
         "concurrently": "^9.2.1",
@@ -27,6 +27,7 @@
     },
     "packageManager": "yarn@4.7.0",
     "resolutions": {
-        "prettier": "3.5.3"
+        "prettier": "3.5.3",
+        "@lezer/common": "1.3.0"
     }
 }

packages/backend/src/index.ts

@@ -94,7 +94,6 @@ const listenToShutdownSignals = () => {
     const cleanup = async (signal: string) => {
         try {
             if (receivedSignal) {
-                logger.debug(`Recieved repeat signal ${signal}, ignoring.`);
                 return;
             }
             receivedSignal = true;

packages/db/src/index.ts

@@ -1 +1,3 @@
+import type { User, Account } from ".prisma/client";
+export type UserWithAccounts = User & { accounts: Account[] };
 export * from ".prisma/client";

packages/mcp/CHANGELOG.md

@@ -7,6 +7,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Changed
+- Updated API client to match the latest Sourcebot release. [#555](https://github.com/sourcebot-dev/sourcebot/pull/555)
+
 ## [1.0.9] - 2025-11-17
 
 ### Added

packages/mcp/src/index.ts

@@ -70,16 +70,12 @@ server.tool(
             query += ` ( lang:${languages.join(' or lang:')} )`;
         }
 
-        if (caseSensitive) {
-            query += ` case:yes`;
-        } else {
-            query += ` case:no`;
-        }
-
         const response = await search({
             query,
             matches: env.DEFAULT_MATCHES,
             contextLines: env.DEFAULT_CONTEXT_LINES,
+            isRegexEnabled: true,
+            isCaseSensitivityEnabled: caseSensitive,
         });
 
         if (isServiceError(response)) {

packages/mcp/src/schemas.ts

@@ -21,15 +21,17 @@ export const symbolSchema = z.object({
     kind: z.string(),
 });
 
+export const searchOptionsSchema = z.object({
+    matches: z.number(),                              // The number of matches to return.
+    contextLines: z.number().optional(),              // The number of context lines to return.
+    whole: z.boolean().optional(),                    // Whether to return the whole file as part of the response.
+    isRegexEnabled: z.boolean().optional(),           // Whether to enable regular expression search.
+    isCaseSensitivityEnabled: z.boolean().optional(), // Whether to enable case sensitivity.
+});
+
 export const searchRequestSchema = z.object({
-    // The zoekt query to execute.
-    query: z.string(),
-    // The number of matches to return.
-    matches: z.number(),
-    // The number of context lines to return.
-    contextLines: z.number().optional(),
-    // Whether to return the whole file as part of the response.
-    whole: z.boolean().optional(),
+    query: z.string(),                                // The zoekt query to execute.
+    ...searchOptionsSchema.shape,
 });
 
 export const repositoryInfoSchema = z.object({
@@ -109,7 +111,7 @@ export const searchStatsSchema = z.object({
     regexpsConsidered: z.number(),
 
     // FlushReason explains why results were flushed.
-    flushReason: z.number(),
+    flushReason: z.string(),
 });
 
 export const searchResponseSchema = z.object({
@@ -139,7 +141,6 @@ export const searchResponseSchema = z.object({
         content: z.string().optional(),
     })),
     repositoryInfo: z.array(repositoryInfoSchema),
-    isBranchFilteringEnabled: z.boolean(),
     isSearchExhaustive: z.boolean(),
 });
 

packages/queryLanguage/.gitignore

@@ -0,0 +1,2 @@
+/node_modules/
+/dist

packages/queryLanguage/package.json

@@ -0,0 +1,20 @@
+{
+    "name": "@sourcebot/query-language",
+    "private": true,
+    "main": "dist/index.js",
+    "scripts": {
+        "build": "lezer-generator src/query.grammar -o src/parser --typeScript --names && tsc",
+        "test": "vitest",
+        "postinstall": "yarn build"
+    },
+    "devDependencies": {
+        "@lezer/generator": "^1.8.0",
+        "tsx": "^4.19.1",
+        "typescript": "^5.7.3",
+        "vitest": "^2.1.9"
+    },
+    "dependencies": {
+        "@lezer/common": "^1.3.0",
+        "@lezer/lr": "^1.4.3"
+    }
+}

packages/queryLanguage/src/index.ts

@@ -0,0 +1,7 @@
+import { parser } from "./parser";
+
+type Tree = ReturnType<typeof parser.parse>;
+type SyntaxNode = Tree['topNode'];
+export type { Tree, SyntaxNode };
+export * from "./parser";
+export * from "./parser.terms";

packages/queryLanguage/src/parser.terms.ts

@@ -0,0 +1,21 @@
+// This file was generated by lezer-generator. You probably shouldn't edit it.
+export const
+  negate = 22,
+  Program = 1,
+  OrExpr = 2,
+  AndExpr = 3,
+  NegateExpr = 4,
+  PrefixExpr = 5,
+  ArchivedExpr = 6,
+  RevisionExpr = 7,
+  ContentExpr = 8,
+  ContextExpr = 9,
+  FileExpr = 10,
+  ForkExpr = 11,
+  VisibilityExpr = 12,
+  RepoExpr = 13,
+  LangExpr = 14,
+  SymExpr = 15,
+  RepoSetExpr = 16,
+  ParenExpr = 17,
+  Term = 18