> ## Documentation Index
> Fetch the complete documentation index at: https://diasporic3lee7-ci-auto-mmdc-diagram-render.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# central-mapper

> Build and query AST indices and dependency graphs with Central Mapper. Rebuild 8,239 indexed files using map-build, map-func, and fore commands.

Central Mapper generates AST indices and dependency graphs for the Termux Monorepo. It parses source files, extracts function signatures, builds file-level graphs, and records cross-module dependencies so you can navigate the codebase by symbol and relationship rather than by directory tree. The indices live in `workspace/llm_map/` and currently cover **8,239 files**.

## Key indices

Central Mapper produces the following indices in `workspace/llm_map/`:

| Index | File | Size | Contents |
| - | - | - | - |
| LLM index | `llm_index.jsonl` | varies | Full AST entries with symbol metadata and source locations |
| Compact LLM index | `llm_index_compact.jsonl` | 1.99 MB | Deduplicated and compressed version of the LLM index |
| Function index | `func_index.jsonl` | 287 KB | Function signatures, parameters, and return types |
| Enriched index | `central_enriched.jsonl` | varies | Enriched records with semantic tags and task associations |
| Task files index | `task_files_index.json` | varies | Mapping from task IDs to the files they touch |
| File graph | `file_graph.json` | varies | Graph of file-level inclusion and import relationships |
| Dependency index | `deps.jsonl` | varies | Cross-module dependency edges with direction and weight |

## Build scripts and rebuild command

Central Mapper provides three build scripts that you chain together to regenerate all indices:

```bash theme={null}
map-build && map-func && fore
```

* **`map-build`** — Scans the repository, parses ASTs, and writes `llm_index.jsonl` and `llm_index_compact.jsonl`
* **`map-func`** — Extracts function-level detail and produces `func_index.jsonl`
* **`fore`** — Runs the forensic pass to build `file_graph.json`, `deps.jsonl`, and `central_enriched.jsonl`

Run the full chain after any significant refactor, merge, or dependency change to keep the indices accurate.

<Warning>
  Rebuilding the full index set on a large repository can take several minutes on an Android device. Run it during low-activity periods or in a background Termux session.
</Warning>

## Directory structure

The `workspace/llm_map/` directory is organized as follows:

```text theme={null}
workspace/llm_map/
├── llm_index.jsonl
├── llm_index_compact.jsonl
├── func_index.jsonl
├── central_enriched.jsonl
├── task_files_index.json
├── file_graph.json
├── deps.jsonl
└── master_tasks.json
```

`master_tasks.json` is the task registry consumed by the Termux Multi-Agent system. Central Mapper updates `task_files_index.json` whenever tasks are modified so the agent orchestrator knows which files each task targets.

## Querying the indices

You can query the indices with standard Unix tools or through ArchWiz forensic modules. For example, to find all functions that reference a specific symbol:

```bash theme={null}
rg "symbol_name" workspace/llm_map/func_index.jsonl
```

To inspect the dependency graph for a file:

```bash theme={null}
python -c "import json; g=json.load(open('workspace/llm_map/file_graph.json')); print(json.dumps(g['src/core.py'], indent=2))"
```

<Note>
  The indices are line-delimited JSON (JSONL) for streaming compatibility. Use `jq` or `rg` for fast queries without loading the entire file into memory.
</Note>

## Integration with ArchWiz

ArchWiz uses Central Mapper indices for forensic and verification workflows. The `correlation_scout.py` and `fragment_matcher.py` tools read `file_graph.json` and `deps.jsonl` to find related files and orphaned fragments. When you run `oracle <file>` from the command line, it consults the dependency index to warn you about downstream impact before you make a change.