### Install Dependencies and Start Development Server

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/README.md

Install project dependencies using npm and start the local development server. Visit http://localhost:3000 to view the site.

```bash
npm install
npm start
```

--------------------------------

### Develop Locally

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/DEPLOYMENT_GUIDE.md

Installs project dependencies and starts the Node.js server for local development. Access the application at http://localhost:3000.

```bash
cd website/
npm install
npm start
# Visit http://localhost:3000
```

--------------------------------

### Big Indexer CLI command for MCP server

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/PHASE9_PLAN.md

Example of starting the Big Indexer MCP server with graph options.

```bash
bgi mcp --graph ...
```

--------------------------------

### Setup for FastAPI Repository Analysis

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_REAL_TRANSCRIPT.md

Initial setup commands to clone the FastAPI repository and prepare for analysis with Big Indexer's MCP tool.

```bash
# Repository: FastAPI (https://github.com/pallets/fastapi)
# Client: GitHub Copilot CLI
# MCP Server: Big Indexer (bgi mcp --graph ... --fuse-graph ...)
# Timestamp: 2026-05-09T19:52:56Z

# Scanned 2,511 units across 14,370 files
# Generated architecture graph with 333 detected clusters
```

--------------------------------

### Example MCP Usage

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/README.md

Illustrates an example usage pattern for the MCP tool, focusing on adding an endpoint and persisting data.

```text
Use MCP tool twin_context for:
"Add endpoint that validates input and persists data."
Return top twin candidate, seam suggestion, and rubric checklist.
```

--------------------------------

### Install Tree-sitter CLI and Language Parser

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/CONTRIBUTING_LANGUAGES.md

Install the tree-sitter command-line interface globally and then install the specific language parser. Replace `<lang>` with the language identifier.

```bash
npm install -g tree-sitter-cli
```

```bash
npm install tree-sitter-<lang>  # e.g., tree-sitter-ruby
```

--------------------------------

### Install and Use Tree-sitter CLI for Ruby

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/LANGUAGE_SUPPORT.md

Install the tree-sitter CLI and the Ruby grammar. Parse a Ruby file to inspect its AST structure and identify key node names.

```bash
npm install -g tree-sitter-cli
npm install tree-sitter-ruby
cat > test.rb << 'EOF'
def fetch_data(url)
  response = HTTPClient.get(url)
  response.body
end
EOF

tree-sitter parse test.rb
```

--------------------------------

### Install and Scan Repository with Big Indexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/blog/I_LEFT_CODY.md

Install the Big Indexer package and then scan a local repository to build a behavioral graph. The `--lang auto` option automatically detects programming languages, and `--out` and `--fuse-graph` specify output file paths.

```bash
pip install bigindexer
cd /path/to/your/big/repo
bgi scan . --lang auto --out bgi-graph.json --fuse-graph fuse-graph.json
```

--------------------------------

### Environment Variables Configuration

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/DEPLOYMENT_GUIDE.md

Sets up essential environment variables for the application. Copy the example file and update the PORT, ADMIN_KEY, and NODE_ENV as needed.

```bash
PORT=3000
ADMIN_KEY=your-secret-key
NODE_ENV=production
```

--------------------------------

### Start MCP Server

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/VALIDATION_EVIDENCE.md

Start the MCP server, which includes BGI-TWIN tools, by providing paths to the BGI graph and fuse graph.

```bash
# Start MCP server (includes BGI-TWIN tools)
bgi mcp --graph output/bgi-graph.json --fuse-graph output/fuse-graph.json
```

--------------------------------

### Start MCP Context Server

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/public/validation.html

Start the MCP context server, which includes BGI-TWIN tools. This server is essential for providing context to AI coding clients.

```bash
# 3. Start the MCP context server (includes BGI-TWIN tools)
bgi mcp --graph output/bgi-graph.json --fuse-graph output/fuse-graph.json
```

--------------------------------

### Troubleshooting: Explicit MCP Tool Reference

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_QUICKSTART_DEMO.md

Example prompts demonstrating the difference between a generic query and an explicit MCP tool reference for better results.

```bash
# Bad (generic):
"Tell me about the architecture"

# Good (explicit):
"Use MCP tool bigindexer_get_architecture_summary to analyze the architecture"
```

--------------------------------

### MCP Tool Invocation Example

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_QUICKSTART_DEMO.md

Example log output showing a call to the 'bigindexer_get_architecture_summary' MCP tool with specific parameters.

```text
CallToolRequest
  Tool: bigindexer_get_architecture_summary
  Parameters: { path_scope: "", top_clusters: 5 }
```

--------------------------------

### Copy Environment Variables

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/README.md

Copy the example environment file to a new file named .env. This file will contain your application's configuration settings.

```bash
cp .env.example .env
```

--------------------------------

### Install and Scan Repository with Bigindexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/blog/I_LEFT_CODY.md

Install the Bigindexer package and then scan a local repository to generate code graph data. Use the `--incremental` flag for subsequent scans on large monorepos to maintain speed.

```bash
pip install bigindexer
cd /your/repo
bgi scan . --lang auto --out bgi-graph.json --fuse-graph fuse-graph.json

# then point your AI client at:
#   bgi mcp --graph bgi-graph.json --fuse-graph fuse-graph.json
```

--------------------------------

### AI Synthesis Example

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_QUICKSTART_DEMO.md

Example of an AI-synthesized summary of Flask's architecture based on MCP tool output.

```text
Flask's architecture centers around three core areas:

• **HTTP Framework**: Request routing, middleware, WSGI integration, 
  response formatting.

• **Templating & Rendering**: Jinja2 integration, template context, 
  auto-escaping for security.

• **CLI & Development Tools**: Development server, debugger, shell context, 
  command registration.
```

--------------------------------

### Verify Big Indexer Installation

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_WITH_CONTINUE.md

Verify the installation by checking the help command for Big Indexer. This confirms the package is accessible in your environment.

```bash
bgi --help
```

--------------------------------

### Install Big Indexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/VALIDATION_EVIDENCE.md

Install the Big Indexer package using pip.

```bash
# Install
pip install bigindexer
```

--------------------------------

### architecture_summary

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/fastapi/opencode_baseline_p04.txt

Get compact architecture summary for context injection.

```APIDOC
## architecture_summary

### Description
Get compact architecture summary for context injection.

### Method
Not applicable (RPC call)

### Endpoint
Not applicable (RPC call)

### Parameters
#### Arguments
- **path_scope** (string) - Optional - The scope of the path for the summary. Defaults to "".
- **top_clusters** (integer) - Optional - The number of top clusters to include. Defaults to 5.
- **seam_limit** (integer) - Optional - The limit for seams to include. Defaults to 10.

### Request Example
```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "architecture_summary",
    "arguments": {
      "path_scope": "fastapi",
      "top_clusters": 10,
      "seam_limit": 20
    }
  }
}
```

### Response
#### Success Response
- **result** (object) - The architecture summary.

#### Response Example
```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "summary": "..."
  }
}
```
```

--------------------------------

### Run MCP Server

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/README.md

Starts the MCP server using generated graph artifacts.

```bash
bgi mcp --graph bgi-graph.json --fuse-graph fuse-graph.json
```

--------------------------------

### Example: Adding Ruby SCM Patterns

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/CONTRIBUTING_LANGUAGES.md

Example SCM query patterns for Ruby, demonstrating how to capture various code constructs including output, emit, transform, mutate, fetch, conditional, loop, exception, guard, async, and fork operations.

```scheme
;; Ruby COV Token Query Patterns

;; OUTPUT: return statements (explicit and implicit)
((return_statement) @output)

;; EMIT: yield statements
((yield_statement) @emit)

;; TRANSFORM: collection methods
((call
  receiver: _ @_recv
  method: (identifier) @name
  (#match? @name "^(map|select|reject|reduce|collect|each_with_index)$"))
 @transform)

;; MUTATE: assignment and mutation methods
((assignment_statement) @mutate)

((call
  receiver: _ @_recv
  method: (identifier) @name
  (#match? @name "^(push|pop|shift|unshift|<<|delete|clear)$"))
 @mutate)

;; FETCH: network and file I/O
((call
  receiver: _ @_recv
  method: (identifier) @name
  (#match? @name "^(get|post|put|delete|open|read|fetch|download)$"))
 @fetch)

;; CONDITIONAL: if/unless/case
((if_statement) @conditional)
((unless_statement) @conditional)
((case_statement) @conditional)

;; LOOP: for/while/loop
((for_statement) @loop)
((while_statement) @loop)
((until_statement) @loop)

;; EXCEPTION: begin/rescue/ensure
((begin_block) @exception)

;; GUARD: type checks
((if_statement
  condition: (call
    method: (identifier) @name
    (#match? @name "^(is_a\?|kind_of\?|instance_of\?)$")))
 @guard)

;; ASYNC: fiber and thread operations
((call
  method: (identifier) @name
  (#match? @name "^(async|await|sleep)$"))
 @async)

;; FORK: threading and multiprocessing
((call
  method: (identifier) @name
  (#match? @name "^(spawn|fork|thread|new)$"))
 @fork)

```

--------------------------------

### Install and Test Bigindexer CLI

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/PHASE9_PLAN.md

Installs the Bigindexer package from PyPI and verifies the command-line interface commands. Ensure the package is published to PyPI before running.

```bash
pip install bigindexer
bgi scan --help
bgi mcp --help
```

--------------------------------

### Install and Scan Repository with BigIndexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/public/blog-reduce-input-token-costs-agentic-tasks.html

Install the BigIndexer package and scan a local repository. Use `--lang auto` for automatic language detection and specify output files for the graph and fuse graph. The `--incremental` flag can be used for subsequent scans on large monorepos to maintain speed.

```bash
pip install bigindexer
cd /your/repo
bgi scan . --lang auto --out bgi-graph.json --fuse-graph fuse-graph.json
```

--------------------------------

### MCP IDE Simulator Command

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/public/demo.html

This command starts the local MCP server for Big Indexer. It loads the graph and binds the registry for JSON-RPC communication.

```bash
ahmedxuhri@bigindexer:~/project$ bgi mcp --graph bgi-graph.json

[BGI] Starting local MCP server v0.1.3... [BGI] Graph loaded: 4,192 units, 8,913 edges, 84 clusters. [BGI] Registry bound: listening for JSON-RPC on stdin.
```

--------------------------------

### Initialize and Query with QueryPlanner

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/QUERY_PLANNER.md

Demonstrates initializing the QueryPlanner with a database path and performing various query types: direct symbol lookup, prefix search, and finding callers. Includes example output structures.

```python
from bgi.indexer.planner import QueryPlanner

# Initialize with index
planner = QueryPlanner(db_path="index.db")

# Query type 1: Direct lookup
results = planner.lookup_symbol("fetch_user", context_unit_id="app.py:main")
# [
#   (Unit(id="auth.py:fetch_user", score=0.92), "Exact match + exported + frequent caller"),
#   (Unit(id="lib.py:fetch_user_cached", score=0.78), "Similar name + local package"),
# ]

# Query type 2: Prefix search
results = planner.search_prefix("fetch", context_unit_id="app.py:main")
# [
#   (Unit(id="auth.py:fetch_user", score=0.92), ...),
#   (Unit(id="auth.py:fetch_data", score=0.81), ...),
#   (Unit(id="cache.py:fetch_cached", score=0.65), ...),
# ]

# Query type 3: Find callers
results = planner.find_callers("fetch_user", max_results=10)
# [
#   (Unit(id="app.py:main", score=0.9), "Direct caller, same file"),
#   (Unit(id="app.py:process", score=0.8), "Indirect via main"),
# ]

# Stats
print(planner.get_stats())
# {
#     "db_path": "index.db",
#     "token_count": 8342,
#     "unique_units": 162954,
#     "packages": 1450,
#     "cache_hit_rate": 0.87,
#     "avg_query_time_ms": 45,
# }
```

--------------------------------

### Get Public Waitlist Status API Response

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/README.md

Example response from the GET /api/waitlist/status endpoint, showing the current size of the waitlist and a status message.

```json
{
  "waitlist_size": 42,
  "message": "People waiting for Big Indexer launch"
}
```

--------------------------------

### Get Full Waitlist API Response (Admin)

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/website/README.md

Example response from the GET /api/admin/waitlist endpoint, which requires an ADMIN_KEY. It returns the total count and a list of emails.

```json
{
  "count": 42,
  "emails": ["user1@example.com", "user2@example.com", ...]
}
```

--------------------------------

### Create Sample Code File

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/CONTRIBUTING_LANGUAGES.md

Create a sample file with representative code for the target language. This file will be used to explore the Abstract Syntax Tree (AST). Replace `<ext>` with the appropriate file extension.

```bash
cat > sample.<ext> << 'EOF'
// Sample code covering:
// - function/method definitions
// - control flow (if, loops)
// - method calls and assignments
// - data operations (returns, yields)
EOF
```

--------------------------------

### Install Big Indexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_WITH_CONTINUE.md

Install the Big Indexer package using pip. Ensure you have Python 3.10+ installed.

```bash
pip install bigindexer
```

--------------------------------

### Install Compatible pydantic-core Version

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/pydantic/opencode_baseline_p01.txt

Use this command to install the specific version of pydantic-core required by your Pydantic installation. This resolves SystemError due to version mismatch.

```bash
pip install 'pydantic-core==2.46.0' 2>&1 | tail -5
```

--------------------------------

### Run MCP Server (stdio)

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_SETUP.md

Start the MCP server from the repository root, providing the generated graph artifacts. Omit --index-db if not using an index database.

```bash
bgi mcp --graph bgi-graph.json --fuse-graph fuse-graph.json --index-db index.db
```

--------------------------------

### Next.js Build Pipeline Overview

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/nextjs/opencode_baseline_p01.txt

Summarizes the steps involved in the `next build` process, including configuration loading, entrypoint creation, bundler selection, and output generation.

```text
next build → build/index.ts
  ├── Load config, env, custom routes
  ├── Create entrypoints (entries.ts)
  ├── parseBundlerArgs() — selects Turbopack (default), Webpack, or Rspack
  │
  ├── [Turbopack] build/turbopack-build/impl.ts
  │     └── Rust NAPI bindings (@next/swc) → native SWC
  │
  ├── [Webpack] build/webpack-build/impl.ts
  │     └── 3 compilers (client + server + edge) with 30+ custom plugins
  │
  ├── Generate manifests (routes, pages, middleware, fonts)
  ├── Prerender static pages
  ├── Build tracing (collect-build-traces.ts — standalone output)
  └── Output to .next/
```

--------------------------------

### Install BigIndexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/README.md

Installs the BigIndexer package in editable mode.

```bash
pip install -e .
```

--------------------------------

### Install and Run Big Indexer via MCP Registry

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/README.md

Install the Big Indexer package and run it using MCP. This command installs version 0.1.3 and generates graph JSON files.

```bash
pip install bigindexer==0.1.3
bgi mcp --graph bgi-graph.json --fuse-graph fuse-graph.json
```

--------------------------------

### Remote Storage Initialization and Fanout

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/prometheus/opencode_mcp_p02_route2.txt

Demonstrates the initialization of remote storage and its integration into a fanout storage mechanism. Remote storage implements the storage.Storage interface and can be coupled with local storage.

```go
remoteStorage = remote.NewStorage(logger, ..., localStorage.StartTime, ...)
fanoutStorage = storage.NewFanout(logger, localStorage, remoteStorage)
```

--------------------------------

### Run with OpenCode

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/VALIDATION_EVIDENCE.md

Run the OpenCode AI assistant with a local configuration file in the repository directory. The AI receives an architecture summary, behavioral twins, seam information, and the scoring rubric.

```bash
# Run with OpenCode and a local config in the repo dir:
# { "mcp": { "bgi": { "command": "bgi", "args": ["mcp", ...] } } }
opencode  # AI receives architecture summary + behavioral twins + seam + rubric
```

--------------------------------

### List Available Tools

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/pydantic/opencode_baseline_p01.txt

Use this command to list all available tools. It sends a JSON-RPC request to the 'tools/list' method.

```bash
echo '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}' | nc -w 3 localhost 8320 2>&1
```

--------------------------------

### MCP Response Example (JSON)

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_QUICKSTART_DEMO.md

Example of a structured JSON response from an MCP tool, detailing repository scope, cluster count, and top clusters.

```json
{
  "scope": "repository",
  "cluster_count": 127,
  "top_clusters": [
    {
      "id": "flask:core",
      "label": "Core HTTP framework",
      "unit_count": 89,
      "description": "Request/response handling, routing, WSGI integration"
    },
    ...
  ]
}
```

--------------------------------

### Run Go Repository Benchmark

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/VALIDATION_EVIDENCE.md

Execute the benchmark script for a Go repository. Ensure the repository and necessary paths are correctly specified.

```bash
# Go repo
python3 scripts/external_benchmark/run_repo.py \
  --repo-slug gin --repo-root /tmp/bgi-ab-repos/gin \
  --package-root . --bgi-graph output/benchmarks/external/gin/bgi-graph.json \
  --ext go --language go \
  --out output/benchmarks/external/gin
```

--------------------------------

### Copilot CLI with Real-time Reasoning for Architecture Summary

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_EXAMPLE_TRANSCRIPTS.md

Demonstrates how the GitHub Copilot CLI uses the MCP bigindexer tool to analyze a repository's architecture and provide a summarized overview. This involves real-time tool invocation during the AI's reasoning process.

```bash
Running the architecture summarizer across the repository to extract the top 
clusters; will use its output to condense the main architectural areas into 
three bullets. Calling tools now.

● architecture_summary (MCP: bigindexer) · path_scope: "", top_clusters: 8, seam_limit: 10
  └ {"scope":"repository","cluster_count":333,"unit_count":2511,"top_cluster...

- **Core framework**: FastAPI runtime (applications, routing, request/response 
handling, datastructures, encoders, OpenAPI generation).
  
- **Dependency, security & concurrency**: dependency-injection utilities, 
background/concurrency helpers, and security (HTTP/OAuth) modules.
  
- **Tests, docs & tooling**: large test suites, documentation examples, and 
scripts (doc/translation tooling and test harness).

Changes   +0 -0
Requests  0 Premium (29s)
Tokens    ↑ 37.4k • ↓ 2.2k • 18.4k (cached) • 2.0k (reasoning)
```

--------------------------------

### Verify Azure CLI Setup

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/DEPLOYMENT_GUIDE.md

Checks if the Azure CLI is configured correctly by displaying the current account information. Ensure you are logged into the correct Azure account.

```bash
az account show
# Should show your Azure account info
```

--------------------------------

### OpenAPI Example Function for Status TSDB Response

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/prometheus/opencode_mcp_p04_route2.txt

Provides example responses for the status TSDB (Time-Series Database) endpoint in the OpenAPI specification. This helps in understanding the structure of successful responses.

```go
func statusTSDBResponseExamples() *orderedmap.Map[string, *base.Example] {
	return orderedmap.NewMap[string, *base.Example]()
}
```

--------------------------------

### Start MCP Server in Background

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/pydantic/opencode_baseline_p01.txt

This command starts the MCP server in the background. It captures the process ID (PID) for later management. Ensure the paths to the graph files are correct for your environment.

```bash
# Start MCP server in background and capture its output
python3 -m bgi.cli mcp --graph /root/mad/sessions/bgi/output/validation/mcp-ab/pydantic/bgi-graph.json --fuse-graph /root/mad/sessions/bgi/output/validation/mcp-ab/pydantic/fuse-graph.json &
MCP_PID=$!
sleep 3
```

--------------------------------

### Fanout Storage Initialization

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/prometheus/opencode_baseline_p02.txt

Illustrates the initialization of a fanout storage layer, combining local and remote storage with specific primary/secondary error handling semantics. This is a critical part of the storage's hierarchical assumption.

```go
fanoutStorage = NewFanout(logger, localStorage, remoteStorage)
```

--------------------------------

### Reproducing FastAPI Analysis with Big Indexer

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/docs/MCP_REAL_TRANSCRIPT.md

Step-by-step commands to reproduce the FastAPI repository analysis, including cloning, scanning, and running the query.

```bash
# Clone FastAPI
git clone https://github.com/pallets/fastapi /tmp/fastapi

# Scan
cd /tmp/fastapi
bgi scan . --lang auto --out bgi-graph.json --fuse-graph fuse-graph.json

# Configure MCP (create ~/.copilot/mcp-config.json with bigindexer server)
# See docs/MCP_SETUP.md for instructions

# Run same query
copilot \
  -C /tmp/fastapi \
  -p "Use MCP tool bigindexer_get_architecture_summary to answer: What are the main architectural clusters in this repo? Provide 3 bullet points." \
  --allow-all
```

--------------------------------

### impact_neighbors

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/fastapi/opencode_baseline_p04.txt

Get likely architectural blast radius from symbol/file.

```APIDOC
## impact_neighbors

### Description
Get likely architectural blast radius from symbol/file.

### Method
Not applicable (RPC call)

### Endpoint
Not applicable (RPC call)

### Parameters
#### Arguments
- **symbol_or_file** (string) - Required - The symbol or file identifier.
- **depth** (integer) - Optional - The depth of the search. Defaults to 2.
- **limit** (integer) - Optional - The maximum number of neighbors to return. Defaults to 50.

### Request Example
```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "impact_neighbors",
    "arguments": {
      "symbol_or_file": "identifier",
      "depth": 3,
      "limit": 100
    }
  }
}
```

### Response
#### Success Response
- **result** (object) - The impact neighbors data.

#### Response Example
```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "neighbors": [
      "neighbor1",
      "neighbor2"
    ]
  }
}
```
```

--------------------------------

### cluster_of_file

Source: https://github.com/ahmedxuhri/bigindexer/blob/master/validation/runs/fastapi/opencode_baseline_p04.txt

Get architectural cluster data for a file path.

```APIDOC
## cluster_of_file

### Description
Get architectural cluster data for a file path.

### Method
Not applicable (RPC call)

### Endpoint
Not applicable (RPC call)

### Parameters
#### Arguments
- **file_path** (string) - Required - The path to the file.

### Request Example
```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "cluster_of_file",
    "arguments": {
      "file_path": "path/to/your/file"
    }
  }
}
```

### Response
#### Success Response
- **result** (object) - The architectural cluster data.

#### Response Example
```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "cluster_data": "..."
  }
}
```
```