Config format V2 (#42)

.gitignore

@@ -1,44 +1,166 @@
-# Created by https://www.toptal.com/developers/gitignore/api/nextjs
-# Edit at https://www.toptal.com/developers/gitignore?templates=nextjs
+# Created by https://www.toptal.com/developers/gitignore/api/yarn,node
+# Edit at https://www.toptal.com/developers/gitignore?templates=yarn,node
 
-### NextJS ###
-# dependencies
-/node_modules
-/.pnp
-.pnp.js
-
-# testing
-/coverage
-
-# next.js
-/.next/
-/out/
-
-# production
-/build
-
-# misc
-.DS_Store
-*.pem
-
-# debug
+### Node ###
+# Logs
+logs
+*.log
 npm-debug.log*
 yarn-debug.log*
 yarn-error.log*
+lerna-debug.log*
 .pnpm-debug.log*
 
-# local env files
-.env*.local
+# Diagnostic reports (https://nodejs.org/api/report.html)
+report.[0-9]*.[0-9]*.[0-9]*.[0-9]*.json
 
-# vercel
-.vercel
+# Runtime data
+pids
+*.pid
+*.seed
+*.pid.lock
 
-# typescript
+# Directory for instrumented libs generated by jscoverage/JSCover
+lib-cov
+
+# Coverage directory used by tools like istanbul
+coverage
+*.lcov
+
+# nyc test coverage
+.nyc_output
+
+# Grunt intermediate storage (https://gruntjs.com/creating-plugins#storing-task-files)
+.grunt
+
+# Bower dependency directory (https://bower.io/)
+bower_components
+
+# node-waf configuration
+.lock-wscript
+
+# Compiled binary addons (https://nodejs.org/api/addons.html)
+build/Release
+
+# Dependency directories
+node_modules/
+jspm_packages/
+
+# Snowpack dependency directory (https://snowpack.dev/)
+web_modules/
+
+# TypeScript cache
 *.tsbuildinfo
-next-env.d.ts
 
-# End of https://www.toptal.com/developers/gitignore/api/nextjs
+# Optional npm cache directory
+.npm
+
+# Optional eslint cache
+.eslintcache
+
+# Optional stylelint cache
+.stylelintcache
+
+# Microbundle cache
+.rpt2_cache/
+.rts2_cache_cjs/
+.rts2_cache_es/
+.rts2_cache_umd/
+
+# Optional REPL history
+.node_repl_history
+
+# Output of 'npm pack'
+*.tgz
+
+# Yarn Integrity file
+.yarn-integrity
+
+# dotenv environment variable files
+.env
+.env.development.local
+.env.test.local
+.env.production.local
+.env.local
+
+# parcel-bundler cache (https://parceljs.org/)
+.cache
+.parcel-cache
+
+# Next.js build output
+.next
+out
+
+# Nuxt.js build / generate output
+.nuxt
+dist
+
+# Gatsby files
+.cache/
+# Comment in the public line in if your project uses Gatsby and not Next.js
+# https://nextjs.org/blog/next-9-1#public-directory-support
+# public
+
+# vuepress build output
+.vuepress/dist
+
+# vuepress v2.x temp and cache directory
+.temp
+
+# Docusaurus cache and generated files
+.docusaurus
+
+# Serverless directories
+.serverless/
+
+# FuseBox cache
+.fusebox/
+
+# DynamoDB Local files
+.dynamodb/
+
+# TernJS port file
+.tern-port
+
+# Stores VSCode versions used for testing VSCode extensions
+.vscode-test
+
+# yarn v2
+.yarn/cache
+.yarn/unplugged
+.yarn/build-state.yml
+.yarn/install-state.gz
+.pnp.*
+
+### Node Patch ###
+# Serverless Webpack directories
+.webpack/
+
+# Optional stylelint cache
+
+# SvelteKit build / generate output
+.svelte-kit
+
+### yarn ###
+# https://yarnpkg.com/getting-started/qa#which-files-should-be-gitignored
+
+.yarn/*
+!.yarn/releases
+!.yarn/patches
+!.yarn/plugins
+!.yarn/sdks
+!.yarn/versions
+
+# if you are NOT using Zero-installs, then:
+# comment the following lines
+!.yarn/cache
+
+# and uncomment the following lines
+# .pnp.*
+
+# End of https://www.toptal.com/developers/gitignore/api/yarn,node
 
 .sourcebot
 /bin
 /config.json
+.DS_Store

.vscode/settings.json

@@ -0,0 +1,6 @@
+{
+    "files.associations": {
+        "*.json": "jsonc",
+        "index.json": "json"
+    }
+}

CHANGELOG.md

@@ -9,6 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- [**Breaking Change**] Added index schema v2. This new schema brings many quality of life features like clearer syntax, ability to specify individual `repos`, `projects`, `groups`, and `orgs`, and the ability to easily `exclude` repositories.
 - Added a `SOURCEBOT_VERSION` build argument to the Docker image. ([#41](https://github.com/sourcebot-dev/sourcebot/pull/41))
 - Added the `sourcebot_version` property to all PostHog events for versioned telemetry. ([#41](https://github.com/sourcebot-dev/sourcebot/pull/41)
 

Dockerfile

@@ -14,18 +14,29 @@ RUN CGO_ENABLED=0 GOOS=linux go build -o /cmd/ ./cmd/...
 FROM node-alpine AS web-builder
 RUN apk add --no-cache libc6-compat
 WORKDIR /app
+
 COPY package.json yarn.lock* ./
+COPY ./packages/web ./packages/web
 
 # Fixes arm64 timeouts
 RUN yarn config set registry https://registry.npmjs.org/
 RUN yarn config set network-timeout 1200000
-RUN yarn --frozen-lockfile
-COPY . .
+RUN yarn workspace @sourcebot/web install --frozen-lockfile
 ENV NEXT_TELEMETRY_DISABLED=1
 # @see: https://phase.dev/blog/nextjs-public-runtime-variables/
 ARG NEXT_PUBLIC_SOURCEBOT_TELEMETRY_DISABLED=BAKED_NEXT_PUBLIC_SOURCEBOT_TELEMETRY_DISABLED
 ARG NEXT_PUBLIC_SOURCEBOT_VERSION=BAKED_NEXT_PUBLIC_SOURCEBOT_VERSION
-RUN yarn run build
+RUN yarn workspace @sourcebot/web build
+
+# ------ Build Backend ------
+FROM node-alpine AS backend-builder
+WORKDIR /app
+
+COPY package.json yarn.lock* ./
+COPY ./schemas ./schemas
+COPY ./packages/backend ./packages/backend
+RUN yarn workspace @sourcebot/backend install --frozen-lockfile
+RUN yarn workspace @sourcebot/backend build
 
 # ------ Runner ------
 FROM node-alpine AS runner
@@ -40,8 +51,8 @@ ARG SOURCEBOT_VERSION=unknown
 ENV SOURCEBOT_VERSION=$SOURCEBOT_VERSION
 RUN echo "Sourcebot Version: $SOURCEBOT_VERSION"
 
-ENV GITHUB_HOSTNAME=github.com
-ENV GITLAB_HOSTNAME=gitlab.com
+# Valid values are: debug, info, warn, error
+ENV SOURCEBOT_LOG_LEVEL=info
 
 # @note: This is also set in .env
 ENV NEXT_PUBLIC_POSTHOG_KEY=phc_VFn4CkEGHRdlVyOOw8mfkoj1DKVoG6y1007EClvzAnS
@@ -50,7 +61,7 @@ ENV NEXT_PUBLIC_POSTHOG_KEY=phc_VFn4CkEGHRdlVyOOw8mfkoj1DKVoG6y1007EClvzAnS
 # ENV SOURCEBOT_TELEMETRY_DISABLED=1
 
 # Configure dependencies
-RUN apk add --no-cache git ca-certificates bind-tools tini jansson wget supervisor uuidgen curl
+RUN apk add --no-cache git ca-certificates bind-tools tini jansson wget supervisor uuidgen curl perl
 
 # Configure zoekt
 COPY vendor/zoekt/install-ctags-alpine.sh .
@@ -68,12 +79,17 @@ COPY --from=zoekt-builder \
     /usr/local/bin/
 
 # Configure the webapp
-COPY --from=web-builder /app/public ./public
-RUN mkdir .next
-COPY --from=web-builder /app/.next/standalone ./
-COPY --from=web-builder /app/.next/static ./.next/static
+COPY --from=web-builder /app/packages/web/public ./packages/web/public
+COPY --from=web-builder /app/packages/web/.next/standalone ./
+COPY --from=web-builder /app/packages/web/.next/static ./packages/web/.next/static
+
+# Configure the backend
+COPY --from=backend-builder /app/node_modules ./node_modules
+COPY --from=backend-builder /app/packages/backend ./packages/backend
 
 COPY supervisord.conf /etc/supervisor/conf.d/supervisord.conf
+COPY prefix-output.sh ./prefix-output.sh
+RUN chmod +x ./prefix-output.sh
 COPY entrypoint.sh ./entrypoint.sh
 RUN chmod +x ./entrypoint.sh
 

Makefile

@@ -9,8 +9,17 @@ ui:
 zoekt:
 	mkdir -p bin
 	go build -C vendor/zoekt -o $(PWD)/bin ./cmd/...
+	export PATH=$(PWD)/bin:$(PATH)
+	export CTAGS_COMMANDS=ctags
 
 clean:
-	rm -rf bin node_modules .next .sourcebot
+	rm -rf \
+		bin \
+		node_modules \
+		packages/web/node_modules \
+		packages/web/.next \
+		packages/backend/dist \
+		packages/backend/node_modules \
+		.sourcebot
 
 .PHONY: bin

README.md

@@ -70,23 +70,25 @@ Sourcebot supports indexing and searching through public and private repositorie
     cd sourcebot_workspace
     ```
 
-2. Create a new config following the [configuration schema](./schemas/index.json) to specify which repositories Sourcebot should index. For example, to index [llama.cpp](https://github.com/ggerganov/llama.cpp):
+2. Create a new config following the [configuration schema](./schemas/v2/index.json) to specify which repositories Sourcebot should index. For example, let's index llama.cpp:
 
     ```sh
     touch my_config.json
     echo '{
-        "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-        "Configs": [
+        "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/refs/tags/latest/schemas/v2/index.json",
+        "repos": [
             {
-                "Type": "github",
-                "GitHubUser": "ggerganov",
-                "Name": "^llama\\\\.cpp$"
+                "type": "github",
+                "repos": [
+                    "ggerganov/llama.cpp"
+                ]
             }
         ]
     }' > my_config.json
     ```
 
-    (For more examples, see [example-config.json](./example-config.json). For additional usage information, see the [configuration schema](./schemas/index.json)).
+>[!NOTE] 
+> Sourcebot can also index all repos owned by a organization, user, group, etc., instead of listing them individually. For examples, see the [configs](./configs) directory. For additional usage information, see the [configuration schema](./schemas/v2/index.json).
 
 3. Run Sourcebot and point it to the new config you created with the `-e CONFIG_PATH` flag:
 
@@ -106,31 +108,8 @@ Sourcebot supports indexing and searching through public and private repositorie
     </details>
     <br>
 
-    You should see a `.sourcebot` folder in your current directory. This folder stores a cache of the repositories zoekt has indexed. The `HEAD` commit of a repository is re-indexed [every hour](https://github.com/sourcebot-dev/zoekt/blob/11b7713f1fb511073c502c41cea413d616f7761f/cmd/zoekt-indexserver/main.go#L86). Indexing private repos? See [Providing an access token](#providing-an-access-token).
+    You should see a `.sourcebot` folder in your current directory. This folder stores a cache of the repositories zoekt has indexed. The `HEAD` commit of a repository is re-indexed [every hour](./packages/backend/src/constants.ts). Indexing private repos? See [Providing an access token](#providing-an-access-token).
     
-    >[!WARNING]
-    > Depending on the size of your repo(s), SourceBot could take a couple of minutes to finish indexing. SourceBot doesn't currently support displaying indexing progress in real-time, so please be patient while it finishes. You can track the progress manually by investigating the `.sourcebot` cache in your workspace.
-
-    <details>
-    <summary><img src="https://gitlab.com/favicon.ico" width="16" height="16" /> Using GitLab?</summary>
-
-    _tl;dr: A `GITLAB_TOKEN` is required to index GitLab repositories (both private & public). See [Providing an access token](#providing-an-access-token)._
-
-    Currently, the GitLab indexer is restricted to only indexing repositories that the associated `GITLAB_TOKEN` has access to. For example, if the token has access to `foo`, `bar`, and `baz` repositories, the following config will index all three:
-
-    ```sh
-    {
-        "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-        "Configs": [
-            {
-                "Type": "gitlab"
-            }
-        ]
-    }
-    ```
-
-    See [Providing an access token](#providing-an-access-token).
-    </details>
     </br>
 
 ## Providing an access token
@@ -145,31 +124,92 @@ This will depend on the code hosting platform you're using:
 </picture> GitHub
 </summary>
 
-In order to index private repositories, you'll need to generate a GitHub Personal Access Token (PAT) and pass it to Sourcebot. Create a new PAT [here](https://github.com/settings/tokens/new) and make sure you select the `repo` scope:
+In order to index private repositories, you'll need to generate a GitHub Personal Access Token (PAT). Create a new PAT [here](https://github.com/settings/tokens/new) and make sure you select the `repo` scope:
 
 ![GitHub PAT creation](.github/images/github-pat-creation.png)
 
-You'll need to pass this PAT each time you run Sourcebot by setting the `GITHUB_TOKEN` environment variable:
+Next, update your configuration with the `token` field:
+```json
+{
+    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/refs/tags/latest/schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "github",
+            "token": "ghp_mytoken",
+            ...
+        }
+    ]
+}
+```
+
+You can also pass tokens as environment variables:
+```json
+{
+    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/refs/tags/latest/schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "github",
+            "token": {
+                // note: this env var can be named anything. It
+                // doesn't need to be `GITHUB_TOKEN`.
+                "env": "GITHUB_TOKEN"
+            },
+            ...
+        }
+    ]
+}
+```
+
+You'll need to pass this environment variable each time you run Sourcebot:
 
 <pre>
-docker run -p 3000:3000 --rm --name sourcebot -e <b>GITHUB_TOKEN=[your-github-token]</b> -e CONFIG_PATH=/data/my_config.json -v $(pwd):/data ghcr.io/sourcebot-dev/sourcebot:latest
+docker run -e <b>GITHUB_TOKEN=ghp_mytoken</b> /* additional args */ ghcr.io/sourcebot-dev/sourcebot:latest
 </pre>
 </details>
 
 <details>
 <summary><img src="https://gitlab.com/favicon.ico" width="16" height="16" /> GitLab</summary>
 
->[!NOTE]
-> An access token is <b>required</b> to index GitLab repositories (both private & public) since the GitLab indexer needs the token to determine which repositories to index. See [example-config.json](./example-config.json) for example usage.
-
 Generate a GitLab Personal Access Token (PAT) [here](https://gitlab.com/-/user_settings/personal_access_tokens) and make sure you select the `read_api` scope:
 
 ![GitLab PAT creation](.github/images/gitlab-pat-creation.png)
 
-You'll need to pass this PAT each time you run Sourcebot by setting the `GITLAB_TOKEN` environment variable:
+Next, update your configuration with the `token` field:
+```json
+{
+    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/refs/tags/latest/schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "gitlab",
+            "token": "glpat-mytoken",
+            ...
+        }
+    ]
+}
+```
+
+You can also pass tokens as environment variables:
+```json
+{
+    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/refs/tags/latest/schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "gitlab",
+            "token": {
+                // note: this env var can be named anything. It
+                // doesn't need to be `GITLAB_TOKEN`.
+                "env": "GITLAB_TOKEN"
+            },
+            ...
+        }
+    ]
+}
+```
+
+You'll need to pass this environment variable each time you run Sourcebot:
 
 <pre>
-docker run -p 3000:3000 --rm --name sourcebot -e <b>GITLAB_TOKEN=[your-gitlab-token]</b> -e CONFIG_PATH=/data/my_config.json -v $(pwd):/data ghcr.io/sourcebot-dev/sourcebot:latest
+docker run -e <b>GITLAB_TOKEN=glpat-mytoken</b> /* additional args */ ghcr.io/sourcebot-dev/sourcebot:latest
 </pre>
 
 </details>
@@ -178,63 +218,7 @@ docker run -p 3000:3000 --rm --name sourcebot -e <b>GITLAB_TOKEN=[your-gitlab-to
 
 ## Using a self-hosted GitLab / GitHub instance
 
-If you're using a self-hosted GitLab or GitHub instance with a custom domain, there is some additional config required:
-
-<div>
-<details>
-<summary>
-<picture>
-    <source media="(prefers-color-scheme: dark)" srcset=".github/images/github-favicon-inverted.png">
-    <img src="https://github.com/favicon.ico" width="16" height="16" alt="GitHub icon">
-</picture> GitHub
-</summary>
-
-1. In your config, add the `GitHubURL` field to point to your deployment's URL. For example:
-    ```json
-    {
-        "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-        "Configs": [
-            {
-                "Type": "github",
-                "GitHubUrl": "https://github.example.com"
-            }
-        ]
-    }
-
-2. Set the `GITHUB_HOSTNAME` environment variable to your deployment's hostname. For example:
-    <pre>
-    docker run -e <b>GITHUB_HOSTNAME=github.example.com</b> /* additional args */ ghcr.io/sourcebot-dev/sourcebot:latest
-    </pre>
-
-
-</details>
-
-<details>
-<summary><img src="https://gitlab.com/favicon.ico" width="16" height="16" /> GitLab</summary>
-
-
-1. In your config, add the `GitLabURL` field to point to your deployment's URL. For example:
-    ```json
-    {
-        "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-        "Configs": [
-            {
-                "Type": "gitlab",
-                "GitLabURL": "https://gitlab.example.com"
-            }
-        ]
-    }
-    ```
-
-2. Set the `GITLAB_HOSTNAME` environment variable to your deployment's hostname. For example:
-
-    <pre>
-    docker run -e <b>GITLAB_HOSTNAME=gitlab.example.com</b> /* additional args */ ghcr.io/sourcebot-dev/sourcebot:latest
-    </pre>
-
-</details>
-
-</div>
+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.
 
 ## Build from source
 >[!NOTE]
@@ -266,49 +250,14 @@ If you're using a self-hosted GitLab or GitHub instance with a custom domain, th
 
 5. Create a `config.json` file at the repository root. See [Configuring Sourcebot](#configuring-sourcebot) for more information.
 
-6. (Optional) Depending on your `config.json`, you may need to pass an access token to Sourcebot:
-
-    <div>
-    <details>
-    <summary>
-    <picture>
-        <source media="(prefers-color-scheme: dark)" srcset=".github/images/github-favicon-inverted.png">
-        <img src="https://github.com/favicon.ico" width="16" height="16" alt="GitHub icon">
-    </picture>
-    GitHub
-    </summary>
-
-    First, generate a personal access token (PAT). See [Providing an access token](#providing-an-access-token).
-
-    Next, Create a text file named `.github-token` **in your home directory** and paste the token in it. The file should look like:
-    ```sh
-    ghp_...
-    ```
-    zoekt will [read this file](https://github.com/sourcebot-dev/zoekt/blob/6a5753692b46e669f851ab23211e756a3677185d/cmd/zoekt-mirror-github/main.go#L60) to authenticate with GitHub.
-    </details>
-
-    <details>
-    <summary>
-        <img src="https://gitlab.com/favicon.ico" width="16" height="16" /> GitLab
-    </summary>
-    First, generate a personal access token (PAT). See [Providing an access token](#providing-an-access-token).
-
-    Next, Create a text file named `.gitlab-token` **in your home directory** and paste the token in it. The file should look like:
-    ```sh
-    glpat-...
-    ```
-    zoekt will [read this file](https://github.com/sourcebot-dev/zoekt/blob/11b7713f1fb511073c502c41cea413d616f7761f/cmd/zoekt-mirror-gitlab/main.go#L43) to authenticate with GitLab.
-    </details>
-    </div>
-
-7. Start Sourcebot with the command:
+6. Start Sourcebot with the command:
     ```sh
     yarn dev
     ```
 
     A `.sourcebot` directory will be created and zoekt will begin to index the repositories found given `config.json`.
 
-8. Start searching at `http://localhost:3000`.
+7. Start searching at `http://localhost:3000`.
 
 ## Telemetry
 

configs/auth.json

@@ -0,0 +1,39 @@
+{
+    "$schema": "../schemas/v2/index.json",
+    "repos": [
+        // Authenticate using a token directly in the config.
+        // Private and public repositories will be included.
+        {
+            "type": "github",
+            "token": "ghp_token1234",
+            "orgs": [
+                "my-org"
+            ]
+        },
+        {
+            "type": "gitlab",
+            "token": "glpat-1234",
+            "groups": [
+                "my-group"
+            ]
+        },
+
+        // You can also store the token in a environment variable and then
+        // references it from the config.
+        {
+            "type": "github",
+            "token": {
+                "env": "GITHUB_TOKEN_ENV_VAR"
+            }
+        },
+        {
+            "type": "gitlab",
+            "token": {
+                "env": "GITLAB_TOKEN_ENV_VAR"
+            },
+            "groups": [
+                "my-group"
+            ]
+        }
+    ]
+}

configs/basic.json

@@ -0,0 +1,42 @@
+{
+    "$schema": "../schemas/v2/index.json",
+    // Note: to include private repositories, you must provide an authentication token.
+    // See: configs/auth.json for a example.
+    "repos": [
+        // From GitHub, include:
+        // - all public repos owned by user `torvalds`
+        // - all public repos owned by organization `commai`
+        // - repo `sourcebot-dev/sourcebot`
+        {
+            "type": "github",
+            "token": "my-token",
+            "users": [
+                "torvalds"
+            ],
+            "orgs": [
+                "commaai"
+            ],
+            "repos": [
+                "sourcebot-dev/sourcebot"
+            ]
+        },
+        // From GitLab, include:
+        // - all public projects owned by user `brendan67`
+        // - all public projects in group `my-group` and sub-group `sub-group`
+        // - project `my-group/project1`
+        {
+            "type": "gitlab",
+            "token": "my-token",
+            "users": [
+                "brendan67"
+            ],
+            "groups": [
+                "my-group",
+                "my-other-group/sub-group"
+            ],
+            "projects": [
+                "my-group/project1"
+            ]
+        }
+    ]
+}

configs/filter.json

@@ -0,0 +1,42 @@
+{
+    "$schema": "../schemas/v2/index.json",
+    "repos": [
+        // Include all repos in my-org, except:
+        // - repo1 & repo2
+        // - repos that are archived or forks
+        {
+            "type": "github",
+            "token": "my-token",
+            "orgs": [
+                "my-org"
+            ],
+            "exclude": {
+                "archived": true,
+                "forks": true,
+                "repos": [
+                    "my-org/repo1",
+                    "my-org/repo2"
+                ]
+            }
+        },
+
+        // Include all projects in my-group, except:
+        // - project1 & project2
+        // - projects that are archived or forks
+        {
+            "type": "gitlab",
+            "token": "my-token",
+            "groups": [
+                "my-group"
+            ],
+            "exclude": {
+                "archived": true,
+                "forks": true,
+                "projects": [
+                    "my-group/project1",
+                    "my-group/project2"
+                ]
+            }
+        }
+    ]
+}

configs/self-hosted.json

@@ -0,0 +1,19 @@
+{
+    "$schema": "../schemas/v2/index.json",
+    "repos": [
+        {
+            "type": "github",
+            "url": "https://github.example.com",
+            "orgs": [
+                "my-org-name"
+            ]
+        },
+        {
+            "type": "gitlab",
+            "url": "https://gitlab.example.com",
+            "groups": [
+                "my-group"
+            ]
+        }
+    ]
+}

default-config.json

@@ -1,10 +1,11 @@
 {
-    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-    "Configs": [
+    "$schema": "./schemas/v2/index.json",
+    "repos": [
         {
-            "Type": "github",
-            "GitHubOrg": "sourcebot-dev",
-            "Name": "^sourcebot$"
+            "type": "github",
+            "repos": [
+                "sourcebot-dev/sourcebot"
+            ]
         }
     ]
 }

demo-site-config.json

@@ -1,41 +1,18 @@
 {
-    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-    "Configs": [
+    "$schema": "./schemas/v2/index.json",
+    "repos": [
         {
-            "Type": "github",
-            "GitHubUser": "torvalds",
-            "Name": "linux"
-        },
-        {
-            "Type": "github",
-            "GitHubOrg": "pytorch",
-            "Name": "pytorch"
-        },
-        {
-            "Type": "github",
-            "GitHubOrg": "commaai",
-            "Name": "^(openpilot|tinygrad)$",
-            "IncludeForks": true
-        },
-        {
-            "Type": "github",
-            "GitHubUser": "ggerganov",
-            "Name": "^(whisper\\.cpp|llama\\.cpp)$"
-        },
-        {
-            "Type": "github",
-            "GitHubOrg": "codemirror",
-            "Name": "^(dev|lang-.*)$"
-        },
-        {
-            "Type": "github",
-            "GitHubOrg": "tailwindlabs",
-            "Name": "^tailwindcss$"
-        },
-        {
-            "Type": "github",
-            "GitHubOrg": "sourcebot-dev",
-            "Name": "^sourcebot$"
+            "type": "github",
+            "repos": [
+                "torvalds/linux",
+                "pytorch/pytorch",
+                "commaai/openpilot",
+                "ggerganov/whisper.cpp",
+                "ggerganov/llama.cpp",
+                "codemirror/dev",
+                "tailwindlabs/tailwindcss",
+                "sourcebot-dev/sourcebot"
+            ]
         }
     ]
 }

entrypoint.sh

@@ -46,36 +46,6 @@ fi
 
 echo -e "\e[34m[Info] Using config file at: '$CONFIG_PATH'.\e[0m"
 
-# Check if GITHUB_TOKEN is set
-if [ -n "$GITHUB_TOKEN" ]; then
-    echo "$GITHUB_TOKEN" > "$HOME/.github-token"
-    chmod 600 "$HOME/.github-token"
-
-    # Configure Git with the provided GITHUB_TOKEN
-    echo -e "\e[34m[Info] Configuring GitHub credentials with hostname '$GITHUB_HOSTNAME'.\e[0m"
-    echo "machine ${GITHUB_HOSTNAME}
-    login oauth
-    password ${GITHUB_TOKEN}" >> "$HOME/.netrc"
-    chmod 600 "$HOME/.netrc"
-else
-    echo -e "\e[34m[Info] Private GitHub repositories will not be indexed since GITHUB_TOKEN was not set.\e[0m"
-fi
-
-# Check if GITLAB_TOKEN is set
-if [ -n "$GITLAB_TOKEN" ]; then
-    echo "$GITLAB_TOKEN" > "$HOME/.gitlab-token"
-    chmod 600 "$HOME/.gitlab-token"
-   
-    # Configure Git with the provided GITLAB_TOKEN
-    echo -e "\e[34m[Info] Configuring GitLab credentials with hostname '$GITLAB_HOSTNAME'.\e[0m"
-    echo "machine ${GITLAB_HOSTNAME}
-    login oauth
-    password ${GITLAB_TOKEN}" >> "$HOME/.netrc"
-    chmod 600 "$HOME/.netrc"
-else
-    echo -e "\e[34m[Info] GitLab repositories will not be indexed since GITLAB_TOKEN was not set.\e[0m"
-fi
-
 # Update nextjs public env variables w/o requiring a rebuild.
 # @see: https://phase.dev/blog/nextjs-public-runtime-variables/
 
@@ -89,7 +59,7 @@ if [ -z "$NEXT_PUBLIC_SOURCEBOT_VERSION" ] && [ ! -z "$SOURCEBOT_VERSION" ]; the
     export NEXT_PUBLIC_SOURCEBOT_VERSION="$SOURCEBOT_VERSION"
 fi
 
-find /app/public /app/.next -type f -name "*.js" |
+find /app/packages/web/public /app/packages/web/.next -type f -name "*.js" |
 while read file; do
     sed -i "s|BAKED_NEXT_PUBLIC_SOURCEBOT_TELEMETRY_DISABLED|${NEXT_PUBLIC_SOURCEBOT_TELEMETRY_DISABLED}|g" "$file"
     sed -i "s|BAKED_NEXT_PUBLIC_SOURCEBOT_VERSION|${NEXT_PUBLIC_SOURCEBOT_VERSION}|g" "$file"

example-config.json

@@ -1,86 +0,0 @@
-{
-    "$schema": "https://raw.githubusercontent.com/sourcebot-dev/sourcebot/main/schemas/index.json",
-    "Configs": [
-        // ~~~~~~~~~~~~ GitHub Examples ~~~~~~~~~~~~
-        // Index all repos in organization "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org"
-        },
-        // Index all repos in self-hosted GitHub instance.
-        // @note: the environment variable GITHUB_HOSTNAME must be set. See README.
-        {
-            "Type": "github",
-            "GitHubUrl": "https://github.example.com"
-        },
-        // Index all repos in user "my-user".
-        {
-            "Type":  "github",
-            "GitHubUser": "my-user"
-        },
-        // Index repos foo & bar in organization "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org",
-            "Name": "^(foo|bar)$"
-        },
-
-        // Index all repos except foo & bar in organization "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org",
-            "Exclude": "^(foo|bar)$"
-        },
-        // Index all repos that contain topic "topic_a" or "topic_b" in organization "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org",
-            "Topics": ["topic_a", "topic_b"]
-        },
-        // Index all repos that _do not_ contain "topic_x" and "topic_y" in organization "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org",
-            "ExcludeTopics": ["topic_x", "topic_y"]
-        },
-        // Index all repos in organization, including forks in "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org",
-            "IncludeForks": true /* default: false */
-        },
-        // Index all repos in organization, excluding repos that are archived in "my-org".
-        {
-            "Type": "github",
-            "GitHubOrg": "my-org",
-            "NoArchived": true /* default: false */
-        }
-
-        // ~~~~~~~~~~~~ GitLab Examples ~~~~~~~~~~~~
-        // Index all repos visible to the GITLAB_TOKEN.
-        {
-            "Type": "gitlab"
-        },
-        // Index all repos visible to the GITLAB_TOKEN (custom GitLab URL).
-        // @note: the environment variable GITLAB_HOSTNAME must also be set. See README.
-        {
-            "Type": "gitlab",
-            "GitLabURL": "https://gitlab.example.com"
-        },
-        // Index all repos (public only) visible to the GITLAB_TOKEN.
-        {
-            "Type": "gitlab",
-            "OnlyPublic": true
-        },
-        // Index only the repos foo & bar.
-        {
-            "Type": "gitlab",
-            "Name": "^(foo|bar)$"
-        },
-        // Index all repos except fizz & buzz visible to the GITLAB_TOKEN.
-        {
-            "Type": "gitlab",
-            "Exclude": "^(fizz|buzz)$"
-        },
-    ]
-}

package.json

@@ -1,83 +1,16 @@
 {
-  "name": "sourcebot",
-  "version": "0.1.0",
     "private": true,
+    "workspaces": [
+        "packages/*"
+    ],
     "scripts": {
-    "dev": "npm-run-all --print-label --parallel next:dev zoekt:webserver zoekt:indexserver",
-    "zoekt:webserver": "export PATH=\"$PWD/bin:$PATH\" && zoekt-webserver -index .sourcebot/index -rpc",
-    "zoekt:indexserver": "export PATH=\"$PWD/bin:$PATH\" && export CTAGS_COMMAND=ctags && zoekt-indexserver -data_dir .sourcebot -mirror_config config.json",
-    "next:dev": "next dev",
-    "build": "next build",
-    "start": "next start",
-    "lint": "next lint"
-  },
-  "dependencies": {
-    "@codemirror/commands": "^6.6.0",
-    "@codemirror/lang-cpp": "^6.0.2",
-    "@codemirror/lang-css": "^6.3.0",
-    "@codemirror/lang-go": "^6.0.1",
-    "@codemirror/lang-html": "^6.4.9",
-    "@codemirror/lang-java": "^6.0.1",
-    "@codemirror/lang-javascript": "^6.2.2",
-    "@codemirror/lang-json": "^6.0.1",
-    "@codemirror/lang-markdown": "^6.2.5",
-    "@codemirror/lang-php": "^6.0.1",
-    "@codemirror/lang-python": "^6.1.6",
-    "@codemirror/lang-rust": "^6.0.1",
-    "@codemirror/lang-sql": "^6.7.1",
-    "@codemirror/search": "^6.5.6",
-    "@codemirror/state": "^6.4.1",
-    "@codemirror/view": "^6.33.0",
-    "@hookform/resolvers": "^3.9.0",
-    "@radix-ui/react-dropdown-menu": "^2.1.1",
-    "@radix-ui/react-icons": "^1.3.0",
-    "@radix-ui/react-label": "^2.1.0",
-    "@radix-ui/react-navigation-menu": "^1.2.0",
-    "@radix-ui/react-scroll-area": "^1.1.0",
-    "@radix-ui/react-separator": "^1.1.0",
-    "@radix-ui/react-slot": "^1.1.0",
-    "@replit/codemirror-lang-csharp": "^6.2.0",
-    "@replit/codemirror-vim": "^6.2.1",
-    "@tanstack/react-query": "^5.53.3",
-    "@tanstack/react-table": "^8.20.5",
-    "@uiw/react-codemirror": "^4.23.0",
-    "class-variance-authority": "^0.7.0",
-    "client-only": "^0.0.1",
-    "clsx": "^2.1.1",
-    "embla-carousel-auto-scroll": "^8.3.0",
-    "embla-carousel-react": "^8.3.0",
-    "escape-string-regexp": "^5.0.0",
-    "http-status-codes": "^2.3.0",
-    "lucide-react": "^0.435.0",
-    "next": "14.2.10",
-    "next-themes": "^0.3.0",
-    "posthog-js": "^1.161.5",
-    "pretty-bytes": "^6.1.1",
-    "react": "^18",
-    "react-dom": "^18",
-    "react-hook-form": "^7.53.0",
-    "react-hotkeys-hook": "^4.5.1",
-    "react-resizable-panels": "^2.1.1",
-    "server-only": "^0.0.1",
-    "sharp": "^0.33.5",
-    "tailwind-merge": "^2.5.2",
-    "tailwindcss-animate": "^1.0.7",
-    "usehooks-ts": "^3.1.0",
-    "zod": "^3.23.8"
+        "build": "yarn workspaces run build",
+        "dev": "npm-run-all --print-label --parallel dev:zoekt dev:backend dev:web",
+        "dev:zoekt": "export PATH=\"$PWD/bin:$PATH\" && zoekt-webserver -index .sourcebot/index -rpc",
+        "dev:backend": "yarn workspace @sourcebot/backend dev:watch",
+        "dev:web": "yarn workspace @sourcebot/web dev"
     },
     "devDependencies": {
-    "@types/node": "^20",
-    "@types/react": "^18",
-    "@types/react-dom": "^18",
-    "@typescript-eslint/eslint-plugin": "^8.3.0",
-    "@typescript-eslint/parser": "^8.3.0",
-    "eslint": "^8",
-    "eslint-config-next": "14.2.6",
-    "eslint-plugin-react": "^7.35.0",
-    "eslint-plugin-react-hooks": "^4.6.2",
-    "npm-run-all": "^4.1.5",
-    "postcss": "^8",
-    "tailwindcss": "^3.4.1",
-    "typescript": "^5"
+        "npm-run-all": "^4.1.5"
     }
 }

packages/backend/.gitignore

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

packages/backend/package.json

@@ -0,0 +1,30 @@
+{
+    "name": "@sourcebot/backend",
+    "version": "0.1.0",
+    "private": true,
+    "main": "index.js",
+    "type": "module",
+    "scripts": {
+        "dev:watch": "yarn generate:types && tsc-watch --preserveWatchOutput --onSuccess \"yarn dev --configPath ../../config.json --cacheDir ../../.sourcebot\"",
+        "dev": "export PATH=\"$PWD/../../bin:$PATH\" && export CTAGS_COMMAND=ctags && node ./dist/index.js",
+        "build": "yarn generate:types && tsc",
+        "generate:types": "tsx tools/generateTypes.ts"
+    },
+    "devDependencies": {
+        "@types/argparse": "^2.0.16",
+        "@types/node": "^22.7.5",
+        "json-schema-to-typescript": "^15.0.2",
+        "tsc-watch": "^6.2.0",
+        "tsx": "^4.19.1",
+        "typescript": "^5.6.2"
+    },
+    "dependencies": {
+        "@gitbeaker/rest": "^40.5.1",
+        "@octokit/rest": "^21.0.2",
+        "argparse": "^2.0.1",
+        "lowdb": "^7.0.1",
+        "simple-git": "^3.27.0",
+        "strip-json-comments": "^5.0.1",
+        "winston": "^3.15.0"
+    }
+}

packages/backend/src/constants.ts

@@ -0,0 +1,10 @@
+
+/**
+ * The interval to reindex a given repository.
+ */
+export const REINDEX_INTERVAL_MS = 1000 * 60 * 60;
+
+/**
+ * The interval to re-sync the config.
+ */
+export const RESYNC_CONFIG_INTERVAL_MS = 1000 * 60 * 60 * 24;

packages/backend/src/db.ts

@@ -0,0 +1,28 @@
+import { JSONFilePreset } from "lowdb/node";
+import { type Low } from "lowdb";
+import { AppContext, Repository } from "./types.js";
+
+type Schema = {
+    repos: {
+        [key: string]: Repository;
+    }
+}
+
+export type Database = Low<Schema>;
+
+export const loadDB = async (ctx: AppContext): Promise<Database> => {
+    const db = await JSONFilePreset<Schema>(`${ctx.cachePath}/db.json`, { repos: {} });
+    return db;
+}
+export const updateRepository = async (repoId: string, data: Partial<Repository>, db: Database) => {
+    db.data.repos[repoId] = {
+        ...db.data.repos[repoId],
+        ...data,
+    }
+    await db.write();
+}
+
+export const createRepository = async (repo: Repository, db: Database) => {
+    db.data.repos[repo.id] = repo;
+    await db.write();
+}

packages/backend/src/environment.ts

@@ -0,0 +1,6 @@
+
+export const getEnv = (env: string | undefined, defaultValue = '') => {
+	return env ?? defaultValue;
+}
+
+export const SOURCEBOT_LOG_LEVEL = getEnv(process.env.SOURCEBOT_LOG_LEVEL, 'info');

packages/backend/src/git.ts

@@ -0,0 +1,51 @@
+import { Repository } from './types.js';
+import { simpleGit, SimpleGitProgressEvent } from 'simple-git';
+import { existsSync } from 'fs';
+import { createLogger } from './logger.js';
+
+const logger = createLogger('git');
+
+export const cloneRepository = async (repo: Repository, onProgress?: (event: SimpleGitProgressEvent) => void) => {
+    if (existsSync(repo.path)) {
+        logger.warn(`${repo.id} already exists. Skipping clone.`)
+        return;
+    }
+
+    const git = simpleGit({
+        progress: onProgress,
+    });
+
+    const gitConfig = Object.entries(repo.gitConfigMetadata ?? {}).flatMap(
+        ([key, value]) => ['--config', `${key}=${value}`]
+    );
+
+    await git.clone(
+        repo.cloneUrl,
+        repo.path,
+        [
+            "--bare",
+            ...gitConfig
+        ]
+    );
+
+    await git.cwd({
+        path: repo.path,
+    }).addConfig("remote.origin.fetch", "+refs/heads/*:refs/heads/*");
+}
+
+
+export const fetchRepository = async (repo: Repository, onProgress?: (event: SimpleGitProgressEvent) => void) => {
+    const git = simpleGit({
+        progress: onProgress,
+    });
+
+    await git.cwd({
+        path: repo.path,
+    }).fetch(
+        "origin",
+        [
+            "--prune",
+            "--progress"
+        ]
+    );
+}

packages/backend/src/github.ts

@@ -0,0 +1,195 @@
+import { Octokit } from "@octokit/rest";
+import { GitHubConfig } from "./schemas/v2.js";
+import { createLogger } from "./logger.js";
+import { AppContext, Repository } from "./types.js";
+import path from 'path';
+import { excludeArchivedRepos, excludeForkedRepos, excludeReposByName, getTokenFromConfig, marshalBool } from "./utils.js";
+
+const logger = createLogger("GitHub");
+
+type OctokitRepository = {
+    name: string,
+    full_name: string,
+    fork: boolean,
+    private: boolean,
+    html_url: string,
+    clone_url?: string,
+    stargazers_count?: number,
+    watchers_count?: number,
+    subscribers_count?: number,
+    forks_count?: number,
+    archived?: boolean,
+}
+
+export const getGitHubReposFromConfig = async (config: GitHubConfig, signal: AbortSignal, ctx: AppContext) => {
+    const token = config.token ? getTokenFromConfig(config.token, ctx) : undefined;
+
+    const octokit = new Octokit({
+        auth: token,
+        ...(config.url ? {
+            baseUrl: `${config.url}/api/v3`
+        } : {}),
+    });
+
+    let allRepos: OctokitRepository[] = [];
+
+    if (config.orgs) {
+        const _repos = await getReposForOrgs(config.orgs, octokit, signal);
+        allRepos = allRepos.concat(_repos);
+    }
+
+    if (config.repos) {
+        const _repos = await getRepos(config.repos, octokit, signal);
+        allRepos = allRepos.concat(_repos);
+    }
+
+    if (config.users) {
+        const isAuthenticated = config.token !== undefined;
+        const _repos = await getReposOwnedByUsers(config.users, isAuthenticated, octokit, signal);
+        allRepos = allRepos.concat(_repos);
+    }
+
+    // Marshall results to our type
+    let repos: Repository[] = allRepos
+        .filter((repo) => {
+            if (!repo.clone_url) {
+                logger.warn(`Repository ${repo.name} missing property 'clone_url'. Excluding.`)
+                return false;
+            }
+            return true;
+        })
+        .map((repo) => {
+            const hostname = config.url ? new URL(config.url).hostname : 'github.com';
+            const repoId = `${hostname}/${repo.full_name}`;
+            const repoPath = path.resolve(path.join(ctx.reposPath, `${repoId}.git`));
+
+            const cloneUrl = new URL(repo.clone_url!);
+            if (token) {
+                cloneUrl.username = token;
+            }
+            
+            return {
+                name: repo.full_name,
+                id: repoId,
+                cloneUrl: cloneUrl.toString(),
+                path: repoPath,
+                isStale: false,
+                isFork: repo.fork,
+                isArchived: !!repo.archived,
+                gitConfigMetadata: {
+                    'zoekt.web-url-type': 'github',
+                    'zoekt.web-url': repo.html_url,
+                    'zoekt.name': repoId,
+                    'zoekt.github-stars': (repo.stargazers_count ?? 0).toString(),
+                    'zoekt.github-watchers': (repo.watchers_count ?? 0).toString(),
+                    'zoekt.github-subscribers': (repo.subscribers_count ?? 0).toString(),
+                    'zoekt.github-forks': (repo.forks_count ?? 0).toString(),
+                    'zoekt.archived': marshalBool(repo.archived),
+                    'zoekt.fork': marshalBool(repo.fork),
+                    'zoekt.public': marshalBool(repo.private === false)
+                }
+            } satisfies Repository;
+        });
+
+    if (config.exclude) {
+        if (!!config.exclude.forks) {
+            repos = excludeForkedRepos(repos, logger);
+        }
+
+        if (!!config.exclude.archived) {
+            repos = excludeArchivedRepos(repos, logger);
+        }
+
+        if (config.exclude.repos) {
+            repos = excludeReposByName(repos, config.exclude.repos, logger);
+        }
+    }
+
+    logger.debug(`Found ${repos.length} total repositories.`);
+    
+    return repos;
+}
+
+const getReposOwnedByUsers = async (users: string[], isAuthenticated: boolean, octokit: Octokit, signal: AbortSignal) => {
+    // @todo : error handling
+    const repos = (await Promise.all(users.map(async (user) => {
+        logger.debug(`Fetching repository info for user ${user}...`);
+        const start = Date.now();
+
+        const result = await (() => {
+            if (isAuthenticated) {
+                return octokit.paginate(octokit.repos.listForAuthenticatedUser, {
+                    username: user,
+                    visibility: 'all',
+                    affiliation: 'owner',
+                    per_page: 100,
+                    request: {
+                        signal,
+                    },
+                });
+            } else {
+                return octokit.paginate(octokit.repos.listForUser, {
+                    username: user,
+                    per_page: 100,
+                    request: {
+                        signal,
+                    },
+                });
+            }
+        })();
+
+        const duration = Date.now() - start;
+        logger.debug(`Found ${result.length} owned by user ${user} in ${duration}ms.`);
+
+        return result;
+    }))).flat();
+
+    return repos;
+}
+
+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();
+
+        const result = await octokit.paginate(octokit.repos.listForOrg, {
+            org: org,
+            per_page: 100,
+            request: {
+                signal
+            }
+        });
+
+        const duration = Date.now() - start;
+        logger.debug(`Found ${result.length} in org ${org} in ${duration}ms.`);
+
+        return result;
+    }))).flat();
+
+    return repos;
+}
+
+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();
+
+        const [owner, repoName] = repo.split('/');
+        const result = await octokit.repos.get({
+            owner,
+            repo: repoName,
+            request: {
+                signal
+            }
+        });
+
+        const duration = Date.now() - start;
+        logger.debug(`Found info for repository ${repo} in ${duration}ms`);
+
+        return result.data;
+    }));
+
+    return repos;
+}

packages/backend/src/gitlab.ts

@@ -0,0 +1,114 @@
+import { Gitlab, ProjectSchema } from "@gitbeaker/rest";
+import { GitLabConfig } from "./schemas/v2.js";
+import { excludeArchivedRepos, excludeForkedRepos, excludeReposByName, getTokenFromConfig, marshalBool, measure } from "./utils.js";
+import { createLogger } from "./logger.js";
+import { AppContext, Repository } from "./types.js";
+import path from 'path';
+
+const logger = createLogger("GitLab");
+
+export const getGitLabReposFromConfig = async (config: GitLabConfig, ctx: AppContext) => {
+    const token = config.token ? getTokenFromConfig(config.token, ctx) : undefined;
+    const api = new Gitlab({
+        ...(config.token ? {
+            token,
+        } : {}),
+        ...(config.url ? {
+            host: config.url,
+        } : {}),
+    });
+
+    let allProjects: ProjectSchema[] = [];
+
+    if (config.groups) {
+        const _projects = (await Promise.all(config.groups.map(async (group) => {
+            logger.debug(`Fetching project info for group ${group}...`);
+            const { durationMs, data } = await measure(() => api.Groups.allProjects(group, {
+                perPage: 100,
+                owned: true,
+            }));
+            logger.debug(`Found ${data.length} projects in group ${group} in ${durationMs}ms.`);
+
+            return data;
+        }))).flat();
+
+        allProjects = allProjects.concat(_projects);
+    }
+
+    if (config.users) {
+        const _projects = (await Promise.all(config.users.map(async (user) => {
+            logger.debug(`Fetching project info for user ${user}...`);
+            const { durationMs, data } = await measure(() => api.Users.allProjects(user, {
+                perPage: 100,
+                owned: true,
+            }));
+            logger.debug(`Found ${data.length} projects owned by user ${user} in ${durationMs}ms.`);
+            return data;
+        }))).flat();
+
+        allProjects = allProjects.concat(_projects);
+    }
+
+    if (config.projects) {
+        const _projects = await Promise.all(config.projects.map(async (project) => {
+            logger.debug(`Fetching project info for project ${project}...`);
+            const { durationMs, data } = await measure(() => api.Projects.show(project));
+            logger.debug(`Found project ${project} in ${durationMs}ms.`);
+            return data;
+        }));
+
+        allProjects = allProjects.concat(_projects);
+    }
+
+    let repos: Repository[] = allProjects
+        .map((project) => {
+            const hostname = config.url ? new URL(config.url).hostname : "gitlab.com";
+            const repoId = `${hostname}/${project.path_with_namespace}`;
+            const repoPath = path.resolve(path.join(ctx.reposPath, `${repoId}.git`))
+            const isFork = project.forked_from_project !== undefined;
+
+            const cloneUrl = new URL(project.http_url_to_repo);
+            if (token) {
+                cloneUrl.username = 'oauth2';
+                cloneUrl.password = token;
+            }
+
+            return {
+                name: project.path_with_namespace,
+                id: repoId,
+                cloneUrl: cloneUrl.toString(),
+                path: repoPath,
+                isStale: false,
+                isFork,
+                isArchived: project.archived,
+                gitConfigMetadata: {
+                    'zoekt.web-url-type': 'gitlab',
+                    'zoekt.web-url': project.web_url,
+                    'zoekt.name': repoId,
+                    'zoekt.gitlab-stars': project.star_count.toString(),
+                    'zoekt.gitlab-forks': project.forks_count.toString(),
+                    'zoekt.archived': marshalBool(project.archived),
+                    'zoekt.fork': marshalBool(isFork),
+                    'zoekt.public': marshalBool(project.visibility === 'public'),
+                }
+            } satisfies Repository;
+        });
+
+    if (config.exclude) {
+        if (!!config.exclude.forks) {
+            repos = excludeForkedRepos(repos, logger);
+        }
+
+        if (!!config.exclude.archived) {
+            repos = excludeArchivedRepos(repos, logger);
+        }
+
+        if (config.exclude.projects) {
+            repos = excludeReposByName(repos, config.exclude.projects, logger);
+        }
+    }
+
+    logger.debug(`Found ${repos.length} total repositories.`);
+
+    return repos;
+}

packages/backend/src/index.ts

@@ -0,0 +1,235 @@
+import { ArgumentParser } from "argparse";
+import { mkdir, readFile } from 'fs/promises';
+import { existsSync, watch } from 'fs';
+import { exec } from "child_process";
+import path from 'path';
+import { SourcebotConfigurationSchema } from "./schemas/v2.js";
+import { getGitHubReposFromConfig } from "./github.js";
+import { getGitLabReposFromConfig } from "./gitlab.js";
+import { AppContext, Repository } from "./types.js";
+import { cloneRepository, fetchRepository } from "./git.js";
+import { createLogger } from "./logger.js";
+import { createRepository, Database, loadDB, updateRepository } from './db.js';
+import { measure } from "./utils.js";
+import { REINDEX_INTERVAL_MS, RESYNC_CONFIG_INTERVAL_MS } from "./constants.js";
+import stripJsonComments from 'strip-json-comments';
+
+const logger = createLogger('main');
+
+const parser = new ArgumentParser({
+    description: "Sourcebot backend tool",
+});
+
+type Arguments = {
+    configPath: string;
+    cacheDir: string;
+}
+
+const indexRepository = async (repo: Repository, ctx: AppContext) => {
+    return new Promise<{ stdout: string, stderr: string }>((resolve, reject) => {
+        exec(`zoekt-git-index -index ${ctx.indexPath} ${repo.path}`, (error, stdout, stderr) => {
+            if (error) {
+                reject(error);
+                return;
+            }
+            resolve({
+                stdout,
+                stderr
+            });
+        })
+    });
+}
+
+const syncConfig = async (configPath: string, db: Database, signal: AbortSignal, ctx: AppContext) => {
+    const configContent = await readFile(configPath, { 
+        encoding: 'utf-8',
+        signal,
+    });
+
+    // @todo: we should validate the configuration file's structure here.
+    const config = JSON.parse(stripJsonComments(configContent)) as SourcebotConfigurationSchema;
+
+    // Fetch all repositories from the config file
+    let configRepos: Repository[] = [];
+    for (const repoConfig of config.repos ?? []) {
+        switch (repoConfig.type) {
+            case 'github': {
+                const gitHubRepos = await getGitHubReposFromConfig(repoConfig, signal, ctx);
+                configRepos.push(...gitHubRepos);
+                break;
+            }
+            case 'gitlab': {
+                const gitLabRepos = await getGitLabReposFromConfig(repoConfig, ctx);
+                configRepos.push(...gitLabRepos);
+                break;
+            }
+        }
+    }
+
+    // De-duplicate on id
+    configRepos.sort((a, b) => {
+        return a.id.localeCompare(b.id);
+    });
+    configRepos = configRepos.filter((item, index, self) => {
+        if (index === 0) return true;
+        if (item.id === self[index - 1].id) {
+            logger.debug(`Duplicate repository ${item.id} found in config file.`);
+            return false;
+        }
+        return true;
+    });
+
+    logger.info(`Discovered ${configRepos.length} unique repositories from config.`);
+
+    // Merge the repositories into the database
+    for (const newRepo of configRepos) {
+        if (newRepo.id in db.data.repos) {
+            await updateRepository(newRepo.id, newRepo, db);
+        } else {
+            await createRepository(newRepo, db);
+        }
+    }
+
+    // Find repositories that are in the database, but not in the configuration file
+    {
+        const a = configRepos.map(repo => repo.id);
+        const b = Object.keys(db.data.repos);
+        const diff = b.filter(x => !a.includes(x));
+
+        for (const id of diff) {
+            await db.update(({ repos }) => {
+                const repo = repos[id];
+                if (repo.isStale) {
+                    return;
+                }
+
+                logger.warn(`Repository ${id} is no longer listed in the configuration file or was not found. Marking as stale.`);
+                repo.isStale = true;
+            });
+        }
+    }
+}
+
+(async () => {
+    parser.add_argument("--configPath", {
+        help: "Path to config file",
+        required: true,
+    });
+
+    parser.add_argument("--cacheDir", {
+        help: "Path to .sourcebot cache directory",
+        required: true,
+    });
+    const args = parser.parse_args() as Arguments;
+
+    if (!existsSync(args.configPath)) {
+        console.error(`Config file ${args.configPath} does not exist`);
+        process.exit(1);
+    }
+
+    const cacheDir = args.cacheDir;
+    const reposPath = path.join(cacheDir, 'repos');
+    const indexPath = path.join(cacheDir, 'index');
+
+    if (!existsSync(reposPath)) {
+        await mkdir(reposPath, { recursive: true });
+    }
+    if (!existsSync(indexPath)) {
+        await mkdir(indexPath, { recursive: true });
+    }
+
+    const context: AppContext = {
+        indexPath,
+        reposPath,
+        cachePath: cacheDir,
+        configPath: args.configPath,
+    }
+
+    const db = await loadDB(context);
+    
+    let abortController = new AbortController();
+    let isSyncing = false;
+    const _syncConfig = () => {
+        if (isSyncing) {
+            abortController.abort();
+            abortController = new AbortController();
+        }
+
+        logger.info(`Syncing configuration file ${args.configPath} ...`);
+        isSyncing = true;
+        measure(() => syncConfig(args.configPath, db, abortController.signal, context))
+            .then(({ durationMs }) => {
+                logger.info(`Synced configuration file ${args.configPath} in ${durationMs / 1000}s`);
+                isSyncing = false;
+            })
+            .catch((err) => {
+                if (err.name === "AbortError") {
+                    // @note: If we're aborting, we don't want to set isSyncing to false
+                    // since it implies another sync is in progress.
+                } else {
+                    isSyncing = false;
+                    logger.error(`Failed to sync configuration file ${args.configPath} with error:\n`, err);
+                }
+            });
+    }
+
+    // Re-sync on file changes
+    watch(args.configPath, () => {
+        logger.info(`Config file ${args.configPath} changed. Re-syncing...`);
+        _syncConfig();
+    });
+
+    // Re-sync every 24 hours
+    setInterval(() => {
+        logger.info(`Re-syncing configuration file ${args.configPath}`);
+        _syncConfig();
+    }, RESYNC_CONFIG_INTERVAL_MS);
+
+    // Sync immediately on startup
+    _syncConfig();
+
+    while (true) {
+        const repos = db.data.repos;
+
+        for (const [_, repo] of Object.entries(repos)) {
+            const lastIndexed = repo.lastIndexedDate ? new Date(repo.lastIndexedDate) : new Date(0);
+
+            if (
+                repo.isStale ||
+                lastIndexed.getTime() > Date.now() - REINDEX_INTERVAL_MS
+            ) {
+                continue;
+            }
+
+            try {
+                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`);
+                }
+
+                logger.info(`Indexing ${repo.id}...`);
+                const { durationMs } = await measure(() => indexRepository(repo, context));
+                logger.info(`Indexed ${repo.id} in ${durationMs / 1000}s`);
+            } catch (err: any) {
+                // @todo : better error handling here..
+                logger.error(err);
+                continue;
+            }
+            
+            await db.update(({ repos }) => repos[repo.id].lastIndexedDate = new Date().toUTCString());
+        }
+
+        await new Promise(resolve => setTimeout(resolve, 1000));
+    }
+})();

packages/backend/src/logger.ts

@@ -0,0 +1,38 @@
+import winston, { format } from 'winston';
+import { SOURCEBOT_LOG_LEVEL } from './environment.js';
+
+const { combine, colorize, timestamp, prettyPrint, errors, printf, label: labelFn } = format;
+
+const createLogger = (label: string) => {
+    return winston.createLogger({
+        // @todo: Make log level configurable
+        level: SOURCEBOT_LOG_LEVEL,
+        format: combine(
+            errors({ stack: true }),
+            timestamp(),
+            prettyPrint(),
+            labelFn({
+                label: label,
+            })
+        ),
+        transports: [
+            new winston.transports.Console({
+                format: combine(
+                    errors({ stack: true }),
+                    colorize(),
+                    printf(({ level, message, timestamp, stack, label: _label }) => {
+                        const label = `[${_label}] `;
+                        if (stack) {
+                            return `${timestamp} ${level}: ${label}${message}\n${stack}`;
+                        }
+                        return `${timestamp} ${level}: ${label}${message}`;
+                    }),
+                ),
+            }),
+        ]
+    });
+}
+
+export {
+    createLogger
+};

packages/backend/src/schemas/v2.ts

@@ -0,0 +1,108 @@
+// THIS IS A AUTO-GENERATED FILE. DO NOT MODIFY MANUALLY!
+
+export type Repos = GitHubConfig | GitLabConfig;
+
+/**
+ * A Sourcebot configuration file outlines which repositories Sourcebot should sync and index.
+ */
+export interface SourcebotConfigurationSchema {
+  $schema?: string;
+  /**
+   * Defines a collection of repositories from varying code hosts that Sourcebot should sync with.
+   */
+  repos?: Repos[];
+}
+export interface GitHubConfig {
+  /**
+   * GitHub Configuration
+   */
+  type: "github";
+  /**
+   * A Personal Access Token (PAT).
+   */
+  token?:
+    | string
+    | {
+        /**
+         * The name of the environment variable that contains the token.
+         */
+        env: string;
+      };
+  /**
+   * The URL of the GitHub host. Defaults to https://github.com
+   */
+  url?: string;
+  /**
+   * List of users to sync with. All repositories that the user owns will be synced, unless explicitly defined in the `exclude` property.
+   */
+  users?: string[];
+  /**
+   * List of organizations to sync with. All repositories in the organization visible to the provided `token` (if any) will be synced, unless explicitly defined in the `exclude` property.
+   */
+  orgs?: string[];
+  /**
+   * List of individual repositories to sync with. Expected to be formatted as '{orgName}/{repoName}' or '{userName}/{repoName}'.
+   */
+  repos?: string[];
+  exclude?: {
+    /**
+     * Exlcude forked repositories from syncing.
+     */
+    forks?: boolean;
+    /**
+     * Exlcude archived repositories from syncing.
+     */
+    archived?: boolean;
+    /**
+     * List of individual repositories to exclude from syncing. Expected to be formatted as '{orgName}/{repoName}' or '{userName}/{repoName}'.
+     */
+    repos?: string[];
+  };
+}
+export interface GitLabConfig {
+  /**
+   * GitLab Configuration
+   */
+  type: "gitlab";
+  /**
+   * An authentication token.
+   */
+  token?:
+    | string
+    | {
+        /**
+         * The name of the environment variable that contains the token.
+         */
+        env: string;
+      };
+  /**
+   * The URL of the GitLab host. Defaults to https://gitlab.com
+   */
+  url?: string;
+  /**
+   * List of users to sync with. All personal projects that the user owns will be synced, unless explicitly defined in the `exclude` property.
+   */
+  users?: string[];
+  /**
+   * List of groups to sync with. All projects in the group visible to the provided `token` (if any) will be synced, unless explicitly defined in the `exclude` property. Subgroups can be specified by providing the path to the subgroup (e.g. `my-group/sub-group-a`).
+   */
+  groups?: string[];
+  /**
+   * List of individual projects to sync with. The project's namespace must be specified. See: https://docs.gitlab.com/ee/user/namespace/
+   */
+  projects?: string[];
+  exclude?: {
+    /**
+     * Exlcude forked projects from syncing.
+     */
+    forks?: boolean;
+    /**
+     * Exlcude archived projects from syncing.
+     */
+    archived?: boolean;
+    /**
+     * List of individual projects to exclude from syncing. The project's namespace must be specified. See: https://docs.gitlab.com/ee/user/namespace/
+     */
+    projects?: string[];
+  };
+}

packages/backend/src/types.ts

@@ -0,0 +1,46 @@
+
+export type Repository = {
+    /**
+     * Name of the repository (e.g., 'sourcebot-dev/sourcebot')
+     */
+    name: string;
+
+    /**
+     * The unique identifier for the repository. (e.g., `github.com/sourcebot-dev/sourcebot`)
+     */
+    id: string;
+    
+    /**
+     * The .git url for the repository
+     */
+    cloneUrl: string;
+
+    /**
+     * Path to where the repository is cloned
+     */
+    path: string;
+
+    gitConfigMetadata?: Record<string, string>;
+
+    lastIndexedDate?: string;
+
+    isStale: boolean;
+    isFork: boolean;
+    isArchived: boolean;
+}
+
+export type AppContext = {
+    /**
+     * Path to the repos cache directory.
+     */
+    reposPath: string;
+
+    /**
+     * Path to the index cache directory;
+     */
+    indexPath: string;
+
+    cachePath: string;
+
+    configPath: string;
+}

packages/backend/src/utils.ts

@@ -0,0 +1,58 @@
+import { Logger } from "winston";
+import { AppContext, Repository } from "./types.js";
+
+export const measure = async <T>(cb : () => Promise<T>) => {
+    const start = Date.now();
+    const data = await cb();
+    const durationMs = Date.now() - start;
+    return {
+        data,
+        durationMs
+    }
+}
+
+export const marshalBool = (value?: boolean) => {
+    return !!value ? '1' : '0';
+}
+
+export const excludeForkedRepos = (repos: Repository[], logger?: Logger) => {
+    return repos.filter((repo) => {
+        if (repo.isFork) {
+            logger?.debug(`Excluding repo ${repo.id}. Reason: exclude.forks is true`);
+            return false;
+        }
+        return true;
+    });
+}
+
+export const excludeArchivedRepos = (repos: Repository[], logger?: Logger) => {
+    return repos.filter((repo) => {
+        if (repo.isArchived) {
+            logger?.debug(`Excluding repo ${repo.id}. Reason: exclude.archived is true`);
+            return false;
+        }
+        return true;
+    });
+}
+
+export const excludeReposByName = (repos: Repository[], excludedRepoNames: string[], logger?: Logger) => {
+    const excludedRepos = new Set(excludedRepoNames);
+    return repos.filter((repo) => {
+        if (excludedRepos.has(repo.name)) {
+            logger?.debug(`Excluding repo ${repo.id}. Reason: exclude.repos contains ${repo.name}`);
+            return false;
+        }
+        return true;
+    });
+}
+
+export const getTokenFromConfig = (token: string | { env: string }, ctx: AppContext) => {
+    if (typeof token === 'string') {
+        return token;
+    }
+    const tokenValue = process.env[token.env];
+    if (!tokenValue) {
+        throw new Error(`The environment variable '${token.env}' was referenced in ${ctx.configPath}, but was not set.`);
+    }
+    return tokenValue;
+}

packages/backend/tools/generateTypes.ts

@@ -0,0 +1,22 @@
+import { compileFromFile } from 'json-schema-to-typescript'
+import path from 'path';
+import fs from 'fs';
+
+const BANNER_COMMENT = '// THIS IS A AUTO-GENERATED FILE. DO NOT MODIFY MANUALLY!\n';
+
+(async () => {
+    const cwd = process.cwd();
+    const schemaPath = path.resolve(`${cwd}/../../schemas/v2/index.json`);
+    const outputPath = path.resolve(`${cwd}/src/schemas/v2.ts`);
+
+    const content = await compileFromFile(schemaPath, {
+        bannerComment: BANNER_COMMENT,
+        cwd,
+    });
+
+    await fs.promises.writeFile(
+        outputPath,
+        content,
+        "utf-8"
+    );
+})();

packages/backend/tsconfig.json

@@ -0,0 +1,26 @@
+{
+	"compilerOptions": {
+		"outDir": "dist",
+		"incremental": true,
+		"declaration": true,
+		"emitDecoratorMetadata": true,
+		"esModuleInterop": true,
+		"experimentalDecorators": true,
+		"forceConsistentCasingInFileNames": true,
+		"isolatedModules": true,
+		"module": "Node16",
+		"moduleResolution": "Node16",
+		"target": "ES2022",
+		"noEmitOnError": false,
+		"noImplicitAny": true,
+		"noUnusedLocals": false,
+		"pretty": true,
+		"resolveJsonModule": true,
+		"skipLibCheck": true,
+		"strict": true,
+		"sourceMap": true,
+        "inlineSources": true
+	},
+	"include": ["src/index.ts"],
+	"exclude": ["node_modules"]
+}

packages/web/.gitignore

@@ -0,0 +1,42 @@
+# Created by https://www.toptal.com/developers/gitignore/api/nextjs
+# Edit at https://www.toptal.com/developers/gitignore?templates=nextjs
+
+### NextJS ###
+# dependencies
+/node_modules
+/.pnp
+.pnp.js
+
+# testing
+/coverage
+
+# next.js
+/.next/
+/out/
+
+# production
+/build
+
+# misc
+.DS_Store
+*.pem
+
+# debug
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+.pnpm-debug.log*
+
+# local env files
+.env*.local
+
+# vercel
+.vercel
+
+# typescript
+*.tsbuildinfo
+next-env.d.ts
+
+# End of https://www.toptal.com/developers/gitignore/api/nextjs
+
+!.env

packages/web/package.json

@@ -0,0 +1,80 @@
+{
+  "name": "@sourcebot/web",
+  "version": "0.1.0",
+  "private": true,
+  "scripts": {
+    "dev": "next dev",
+    "build": "next build",
+    "start": "next start",
+    "lint": "next lint"
+  },
+  "dependencies": {
+    "@codemirror/commands": "^6.6.0",
+    "@codemirror/lang-cpp": "^6.0.2",
+    "@codemirror/lang-css": "^6.3.0",
+    "@codemirror/lang-go": "^6.0.1",
+    "@codemirror/lang-html": "^6.4.9",
+    "@codemirror/lang-java": "^6.0.1",
+    "@codemirror/lang-javascript": "^6.2.2",
+    "@codemirror/lang-json": "^6.0.1",
+    "@codemirror/lang-markdown": "^6.2.5",
+    "@codemirror/lang-php": "^6.0.1",
+    "@codemirror/lang-python": "^6.1.6",
+    "@codemirror/lang-rust": "^6.0.1",
+    "@codemirror/lang-sql": "^6.7.1",
+    "@codemirror/search": "^6.5.6",
+    "@codemirror/state": "^6.4.1",
+    "@codemirror/view": "^6.33.0",
+    "@hookform/resolvers": "^3.9.0",
+    "@radix-ui/react-dropdown-menu": "^2.1.1",
+    "@radix-ui/react-icons": "^1.3.0",
+    "@radix-ui/react-label": "^2.1.0",
+    "@radix-ui/react-navigation-menu": "^1.2.0",
+    "@radix-ui/react-scroll-area": "^1.1.0",
+    "@radix-ui/react-separator": "^1.1.0",
+    "@radix-ui/react-slot": "^1.1.0",
+    "@replit/codemirror-lang-csharp": "^6.2.0",
+    "@replit/codemirror-vim": "^6.2.1",
+    "@tanstack/react-query": "^5.53.3",
+    "@tanstack/react-table": "^8.20.5",
+    "@uiw/react-codemirror": "^4.23.0",
+    "class-variance-authority": "^0.7.0",
+    "client-only": "^0.0.1",
+    "clsx": "^2.1.1",
+    "embla-carousel-auto-scroll": "^8.3.0",
+    "embla-carousel-react": "^8.3.0",
+    "escape-string-regexp": "^5.0.0",
+    "http-status-codes": "^2.3.0",
+    "lucide-react": "^0.435.0",
+    "next": "14.2.10",
+    "next-themes": "^0.3.0",
+    "posthog-js": "^1.161.5",
+    "pretty-bytes": "^6.1.1",
+    "react": "^18",
+    "react-dom": "^18",
+    "react-hook-form": "^7.53.0",
+    "react-hotkeys-hook": "^4.5.1",
+    "react-resizable-panels": "^2.1.1",
+    "server-only": "^0.0.1",
+    "sharp": "^0.33.5",
+    "tailwind-merge": "^2.5.2",
+    "tailwindcss-animate": "^1.0.7",
+    "usehooks-ts": "^3.1.0",
+    "zod": "^3.23.8"
+  },
+  "devDependencies": {
+    "@types/node": "^20",
+    "@types/react": "^18",
+    "@types/react-dom": "^18",
+    "@typescript-eslint/eslint-plugin": "^8.3.0",
+    "@typescript-eslint/parser": "^8.3.0",
+    "eslint": "^8",
+    "eslint-config-next": "14.2.6",
+    "eslint-plugin-react": "^7.35.0",
+    "eslint-plugin-react-hooks": "^4.6.2",
+    "npm-run-all": "^4.1.5",
+    "postcss": "^8",
+    "tailwindcss": "^3.4.1",
+    "typescript": "^5"
+  }
+}