Metadata-Version: 2.4
Name: onetick-llmbase
Version: 0.30.0b2
Summary: MCP server engine and indexing library for OneTick LLM assistants — semantic document search and reference-data tools; configured and launched by a wrapper package (e.g. onetick-mcp).
Author-email: solutions <solutions@onetick.com>
License-Expression: MIT
Project-URL: onetick-mcp docs, https://onetickmcp.docs.sol.onetick.com
Project-URL: OneTick, https://onetick.com
Requires-Python: <3.15,>=3.12
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: alembic<2,>=1.13.1
Requires-Dist: click
Requires-Dist: fastmcp<4,>=3
Requires-Dist: httpx
Requires-Dist: langchain-classic<2,>=1.0.1
Requires-Dist: langchain-community<1,>=0.4.1
Requires-Dist: langchain-core<2,>=1.2.7
Requires-Dist: langchain-openai<2,>=1.1.7
Requires-Dist: langchain-postgres<0.1,>=0.0.16
Requires-Dist: langchain-text-splitters<2,>=1.1.0
Requires-Dist: more_itertools<11,>=10.6.0
Requires-Dist: onetick-py[webapi]<2
Requires-Dist: pandas
Requires-Dist: psycopg[binary,pool]
Requires-Dist: pydantic-settings<3.0.0,>=2.0.0
Requires-Dist: python-slugify<9,>=8.0.4
Requires-Dist: SQLAlchemy<2.1,>=2.0.31
Requires-Dist: tiktoken<1,>=0.8
Requires-Dist: tqdm
Requires-Dist: urllib3>=2.7
Provides-Extra: local-embeddings
Requires-Dist: sentence-transformers>=2.7.0; extra == "local-embeddings"
Requires-Dist: langchain-huggingface<2,>=1.2.0; extra == "local-embeddings"
Provides-Extra: translator
Requires-Dist: transformers<5,>=4.43.3; extra == "translator"
Requires-Dist: sentencepiece<1,>=0.2.0; extra == "translator"
Requires-Dist: langid<2,>=1.1.6; extra == "translator"
Provides-Extra: dev
Requires-Dist: pytest<9; extra == "dev"
Requires-Dist: pytest-cov<6; extra == "dev"
Requires-Dist: ruff; extra == "dev"
Dynamic: license-file

# onetick-llmbase

`onetick-llmbase` is the library behind [`onetick-mcp`](https://pypi.org/project/onetick-mcp/) — the Model Context Protocol server for OneTick. It provides the retrieval/tool engine, the pgvector document-indexing pipeline, and the three tool categories the server exposes — Documentation search, Reference data, and OneTick queries.

Most users should install `onetick-mcp` directly. Install `onetick-llmbase` separately only if you are building a custom MCP server on top of this engine.

## Documentation

Full documentation is part of the [onetick-mcp package documentation](https://onetickmcp.docs.sol.onetick.com).

## Dependencies

- **MCP / server**: `fastmcp`, `starlette`, `click`, `httpx`, `pydantic-settings`
- **Search & indexing**: `langchain-core`, `langchain-classic`, `langchain-community`, `langchain-openai`,
  `langchain-postgres`, `langchain-text-splitters`, `tiktoken`
- **Data**: `SQLAlchemy`, `psycopg`, `alembic`, `pandas`
- **OneTick access**: `onetick-py[webapi]`
- **Utilities**: `more_itertools`, `python-slugify`, `tqdm`, `urllib3`

Optional extras:

| Extra | Installs | When to use |
|---|---|---|
| `local-embeddings` | `sentence-transformers`, `langchain-huggingface` | On-prem deployments with no external egress, or users who don't want to use OpenAI; set `EMBEDDING_TYPE=sentence_transformer` |

## Install

```bash
pip install onetick-llmbase
```
