rag-python 0.3.0__tar.gz → 0.3.1__tar.gz

rag_python-0.3.1/PKG-INFO

@@ -0,0 +1,205 @@
+Metadata-Version: 2.2
+Name: rag-python
+Version: 0.3.1
+Summary: Production-grade RAG for Python: ingest documents, hybrid search, streaming, multi-LLM, guardrails. CLI + Python API.
+Author-email: Raghav Singla <04raghavsingla28@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/RaghavOG/rag-python
+Project-URL: Repository, https://github.com/RaghavOG/rag-python
+Project-URL: Documentation, https://github.com/RaghavOG/rag-python/tree/main/docs
+Project-URL: Changelog, https://github.com/RaghavOG/rag-python/blob/main/CHANGELOG.md
+Project-URL: Issues, https://github.com/RaghavOG/rag-python/issues
+Keywords: rag,llm,embeddings,chromadb,openai,rag-python,retrieval-augmented-generation
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: openai>=1.12.0
+Requires-Dist: tiktoken>=0.5.0
+Requires-Dist: chromadb>=0.4.22
+Requires-Dist: pypdf>=3.17.0
+Requires-Dist: python-docx>=1.1.0
+Requires-Dist: langdetect>=1.0.9
+Requires-Dist: regex>=2023.0.0
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: requests>=2.31.0
+Provides-Extra: rerank
+Requires-Dist: sentence-transformers>=2.2.0; extra == "rerank"
+Requires-Dist: torch>=2.0.0; extra == "rerank"
+Provides-Extra: local
+Requires-Dist: sentence-transformers>=2.2.0; extra == "local"
+Provides-Extra: hybrid
+Requires-Dist: rank-bm25>=0.2.2; extra == "hybrid"
+Provides-Extra: anthropic
+Requires-Dist: anthropic>=0.20.0; extra == "anthropic"
+Provides-Extra: gemini
+Requires-Dist: google-genai>=0.3.0; extra == "gemini"
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: ruff>=0.1.0; extra == "dev"
+Requires-Dist: build; extra == "dev"
+Requires-Dist: twine; extra == "dev"
+Requires-Dist: rank-bm25>=0.2.2; extra == "dev"
+Provides-Extra: all
+Requires-Dist: rag-python[anthropic,gemini,hybrid,local,rerank]; extra == "all"
+
+# rag-python
+
+[![PyPI version](https://img.shields.io/pypi/v/rag-python.svg)](https://pypi.org/project/rag-python/)
+[![PyPI downloads](https://img.shields.io/pypi/dm/rag-python.svg)](https://pypi.org/project/rag-python/)
+[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![Documentation](https://img.shields.io/badge/docs-GitHub-blue)](https://github.com/RaghavOG/rag-python/tree/main/docs)
+
+**Production-grade Retrieval-Augmented Generation (RAG) for Python** — ingest documents, ask questions, get grounded answers with multi-LLM support, hybrid search, streaming, and guardrails.
+
+```bash
+pip install rag-python
+export OPENAI_API_KEY=sk-...
+rag-python ingest ./docs --reindex
+rag-python query "What is our leave policy?"
+```
+
+**Author:** [Raghav Singla](https://github.com/RaghavOG) · **Repo:** [github.com/RaghavOG/rag-python](https://github.com/RaghavOG/rag-python)
+
+---
+
+## Why rag-python?
+
+| Capability | What you get |
+|------------|--------------|
+| **Ingest** | TXT, MD, PDF, DOCX, CSV, JSON, HTML → chunk → embed → ChromaDB |
+| **Retrieve** | Multi-query rewriting, **hybrid BM25+vector**, reranking, metadata filters |
+| **Generate** | Multi-LLM answers with guardrails, evaluation, and retry loop |
+| **Stream** | `rag.query_stream()` and `--stream` CLI for responsive UX |
+| **Offline** | Local embeddings via sentence-transformers |
+| **CLI** | `rag-python ingest`, `query`, `docs` — no code required |
+
+---
+
+## Install
+
+```bash
+pip install rag-python
+```
+
+| Extra | Install | Enables |
+|-------|---------|---------|
+| `local` | `pip install rag-python[local]` | Offline embeddings (sentence-transformers) |
+| `hybrid` | `pip install rag-python[hybrid]` | BM25 + vector hybrid retrieval |
+| `rerank` | `pip install rag-python[rerank]` | Cross-encoder reranking |
+| `anthropic` | `pip install rag-python[anthropic]` | Claude LLM |
+| `gemini` | `pip install rag-python[gemini]` | Gemini LLM |
+| `all` | `pip install rag-python[all]` | All optional features |
+
+---
+
+## Quickstart (Python)
+
+```python
+from rag_python import RAG
+
+rag = RAG(llm_model="gpt-4o-mini")
+rag.ingest(["./data"], reindex=True)
+
+answer = rag.query("How many days of annual leave?")
+print(answer.text)
+print(answer.sources)
+```
+
+### Streaming
+
+```python
+stream = rag.query_stream("How many days of annual leave?")
+for token in stream:
+    print(token, end="", flush=True)
+print(stream.result.evaluation)
+```
+
+### Hybrid search + metadata filter
+
+```python
+rag = RAG(
+    retriever="hybrid",  # pip install rag-python[hybrid]
+    metadata_filter={"filename": "leave-policy.pdf"},
+)
+rag.ingest(["./policies/"])
+print(rag.query("annual leave policy").text)
+```
+
+---
+
+## Quickstart (CLI)
+
+```bash
+export OPENAI_API_KEY=sk-...
+
+rag-python ingest ./data --reindex
+rag-python query "How many days of annual leave?"
+rag-python query "PTO policy" --stream -v
+rag-python query "benefits" --retriever hybrid
+
+# Built-in terminal docs
+rag-python docs quickstart
+rag-python docs --list
+rag-python --help
+```
+
+---
+
+## Documentation
+
+| Guide | Description |
+|-------|-------------|
+| [**Docs index**](docs/README.md) | Start here |
+| [Usage](docs/USAGE.md) | Python API, streaming, retrieval |
+| [CLI reference](docs/CLI.md) | All `rag-python` commands and flags |
+| [Configuration](docs/CONFIGURATION.md) | Env vars and `RAGConfig` |
+| [Providers](docs/PROVIDERS.md) | OpenAI, Azure, Anthropic, Gemini, Ollama, local |
+| [Changelog](CHANGELOG.md) | Release notes |
+
+**In the terminal:** `rag-python docs [topic]` — topics: `quickstart`, `install`, `cli`, `config`, `providers`, `features`
+
+---
+
+## Environment variables
+
+| Variable | Description |
+|----------|-------------|
+| `OPENAI_API_KEY` | Default LLM + embeddings |
+| `ANTHROPIC_API_KEY` | Claude |
+| `GEMINI_API_KEY` | Gemini |
+| `AZURE_OPENAI_ENDPOINT` / `AZURE_OPENAI_API_KEY` | Azure OpenAI |
+| `OLLAMA_BASE_URL` | Local Ollama (default `http://localhost:11434`) |
+| `RAG_PYTHON_DATA_DIR` | Document dir (default `./data`) |
+| `RAG_PYTHON_CHROMA_DIR` | Vector store (default `./chroma_db`) |
+
+See [Configuration](docs/CONFIGURATION.md) and [`.env.example`](.env.example).
+
+---
+
+## Project layout
+
+```text
+src/rag_python/     # pip install rag-python → import rag_python
+  client.py         # RAG, RAGAnswer, query_stream
+  rag_pipeline.py   # ingest / query pipeline
+  providers/        # OpenAI, Azure, Anthropic, Gemini, Ollama, local
+docs/               # User documentation (linked from PyPI README)
+tests/
+examples/
+```
+
+---
+
+## License
+
+MIT © [Raghav Singla](https://github.com/RaghavOG)

rag_python-0.3.1/README.md

@@ -0,0 +1,152 @@
+# rag-python
+
+[![PyPI version](https://img.shields.io/pypi/v/rag-python.svg)](https://pypi.org/project/rag-python/)
+[![PyPI downloads](https://img.shields.io/pypi/dm/rag-python.svg)](https://pypi.org/project/rag-python/)
+[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![Documentation](https://img.shields.io/badge/docs-GitHub-blue)](https://github.com/RaghavOG/rag-python/tree/main/docs)
+
+**Production-grade Retrieval-Augmented Generation (RAG) for Python** — ingest documents, ask questions, get grounded answers with multi-LLM support, hybrid search, streaming, and guardrails.
+
+```bash
+pip install rag-python
+export OPENAI_API_KEY=sk-...
+rag-python ingest ./docs --reindex
+rag-python query "What is our leave policy?"
+```
+
+**Author:** [Raghav Singla](https://github.com/RaghavOG) · **Repo:** [github.com/RaghavOG/rag-python](https://github.com/RaghavOG/rag-python)
+
+---
+
+## Why rag-python?
+
+| Capability | What you get |
+|------------|--------------|
+| **Ingest** | TXT, MD, PDF, DOCX, CSV, JSON, HTML → chunk → embed → ChromaDB |
+| **Retrieve** | Multi-query rewriting, **hybrid BM25+vector**, reranking, metadata filters |
+| **Generate** | Multi-LLM answers with guardrails, evaluation, and retry loop |
+| **Stream** | `rag.query_stream()` and `--stream` CLI for responsive UX |
+| **Offline** | Local embeddings via sentence-transformers |
+| **CLI** | `rag-python ingest`, `query`, `docs` — no code required |
+
+---
+
+## Install
+
+```bash
+pip install rag-python
+```
+
+| Extra | Install | Enables |
+|-------|---------|---------|
+| `local` | `pip install rag-python[local]` | Offline embeddings (sentence-transformers) |
+| `hybrid` | `pip install rag-python[hybrid]` | BM25 + vector hybrid retrieval |
+| `rerank` | `pip install rag-python[rerank]` | Cross-encoder reranking |
+| `anthropic` | `pip install rag-python[anthropic]` | Claude LLM |
+| `gemini` | `pip install rag-python[gemini]` | Gemini LLM |
+| `all` | `pip install rag-python[all]` | All optional features |
+
+---
+
+## Quickstart (Python)
+
+```python
+from rag_python import RAG
+
+rag = RAG(llm_model="gpt-4o-mini")
+rag.ingest(["./data"], reindex=True)
+
+answer = rag.query("How many days of annual leave?")
+print(answer.text)
+print(answer.sources)
+```
+
+### Streaming
+
+```python
+stream = rag.query_stream("How many days of annual leave?")
+for token in stream:
+    print(token, end="", flush=True)
+print(stream.result.evaluation)
+```
+
+### Hybrid search + metadata filter
+
+```python
+rag = RAG(
+    retriever="hybrid",  # pip install rag-python[hybrid]
+    metadata_filter={"filename": "leave-policy.pdf"},
+)
+rag.ingest(["./policies/"])
+print(rag.query("annual leave policy").text)
+```
+
+---
+
+## Quickstart (CLI)
+
+```bash
+export OPENAI_API_KEY=sk-...
+
+rag-python ingest ./data --reindex
+rag-python query "How many days of annual leave?"
+rag-python query "PTO policy" --stream -v
+rag-python query "benefits" --retriever hybrid
+
+# Built-in terminal docs
+rag-python docs quickstart
+rag-python docs --list
+rag-python --help
+```
+
+---
+
+## Documentation
+
+| Guide | Description |
+|-------|-------------|
+| [**Docs index**](docs/README.md) | Start here |
+| [Usage](docs/USAGE.md) | Python API, streaming, retrieval |
+| [CLI reference](docs/CLI.md) | All `rag-python` commands and flags |
+| [Configuration](docs/CONFIGURATION.md) | Env vars and `RAGConfig` |
+| [Providers](docs/PROVIDERS.md) | OpenAI, Azure, Anthropic, Gemini, Ollama, local |
+| [Changelog](CHANGELOG.md) | Release notes |
+
+**In the terminal:** `rag-python docs [topic]` — topics: `quickstart`, `install`, `cli`, `config`, `providers`, `features`
+
+---
+
+## Environment variables
+
+| Variable | Description |
+|----------|-------------|
+| `OPENAI_API_KEY` | Default LLM + embeddings |
+| `ANTHROPIC_API_KEY` | Claude |
+| `GEMINI_API_KEY` | Gemini |
+| `AZURE_OPENAI_ENDPOINT` / `AZURE_OPENAI_API_KEY` | Azure OpenAI |
+| `OLLAMA_BASE_URL` | Local Ollama (default `http://localhost:11434`) |
+| `RAG_PYTHON_DATA_DIR` | Document dir (default `./data`) |
+| `RAG_PYTHON_CHROMA_DIR` | Vector store (default `./chroma_db`) |
+
+See [Configuration](docs/CONFIGURATION.md) and [`.env.example`](.env.example).
+
+---
+
+## Project layout
+
+```text
+src/rag_python/     # pip install rag-python → import rag_python
+  client.py         # RAG, RAGAnswer, query_stream
+  rag_pipeline.py   # ingest / query pipeline
+  providers/        # OpenAI, Azure, Anthropic, Gemini, Ollama, local
+docs/               # User documentation (linked from PyPI README)
+tests/
+examples/
+```
+
+---
+
+## License
+
+MIT © [Raghav Singla](https://github.com/RaghavOG)

{rag_python-0.3.0 → rag_python-0.3.1}/pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rag-python"
-version = "0.3.0"
-description = "Production-grade RAG for Python: multi-LLM, query rewriting, reranking, guardrails, and evaluation."
+version = "0.3.1"
+description = "Production-grade RAG for Python: ingest documents, hybrid search, streaming, multi-LLM, guardrails. CLI + Python API."
 readme = "README.md"
 license = { text = "MIT" }
 requires-python = ">=3.10"
@@ -51,7 +51,8 @@ rag-python = "rag_python.cli:main"
 [project.urls]
 Homepage = "https://github.com/RaghavOG/rag-python"
 Repository = "https://github.com/RaghavOG/rag-python"
-Documentation = "https://github.com/RaghavOG/rag-python#readme"
+Documentation = "https://github.com/RaghavOG/rag-python/tree/main/docs"
+Changelog = "https://github.com/RaghavOG/rag-python/blob/main/CHANGELOG.md"
 Issues = "https://github.com/RaghavOG/rag-python/issues"
 
 [tool.setuptools.packages.find]

{rag_python-0.3.0 → rag_python-0.3.1}/src/rag_python/__init__.py

@@ -2,18 +2,29 @@
 
 Quick start::
 
-    from rag_python import RAG
+    pip install rag-python
+    export OPENAI_API_KEY=sk-...
+
+    # CLI
+    rag-python ingest ./docs --reindex
+    rag-python query "What is our leave policy?"
+    rag-python docs quickstart
 
-    rag = RAG(llm_model="gpt-4o-mini")
+    # Python
+    from rag_python import RAG
+    rag = RAG()
     rag.ingest(["./docs"], reindex=True)
     print(rag.query("What is our leave policy?").text)
+
+Documentation: https://github.com/RaghavOG/rag-python/tree/main/docs
 """
 
-__version__ = "0.3.0"
+__version__ = "0.3.1"
 
 from .client import RAG, RAGAnswer
-from .rag_pipeline import ingest, query, RAGResponse
+from .rag_pipeline import ingest, query, query_stream, RAGResponse, RAGStream
 from .providers import make_llm_provider, make_embedding_provider
+from .log import configure_logging, get_logger
 from .options import (
     ChunkingConfig,
     DocumentConfig,
@@ -33,7 +44,11 @@ __all__ = [
     "QueryConfig",
     "ingest",
     "query",
+    "query_stream",
+    "RAGStream",
     "RAGResponse",
+    "configure_logging",
+    "get_logger",
     "make_llm_provider",
     "make_embedding_provider",
 ]

rag_python-0.3.1/src/rag_python/cli.py

@@ -0,0 +1,266 @@
+"""rag-python command-line interface."""
+from __future__ import annotations
+
+import argparse
+import json
+import sys
+from dataclasses import replace
+
+from . import __version__
+from .client import RAG
+from .help_text import CLI_EPILOG, list_topics, print_topic, print_topic_list
+
+
+def _build_rag(args: argparse.Namespace) -> RAG:
+    kwargs: dict = dict(
+        llm_provider=args.llm_provider,
+        llm_model=args.llm_model,
+        embedding_provider=args.embedding_provider,
+        embedding_model=args.embedding_model,
+        openai_api_key=args.openai_api_key,
+        azure_endpoint=args.azure_endpoint,
+        azure_api_key=args.azure_api_key,
+        azure_api_version=args.azure_api_version,
+        anthropic_api_key=args.anthropic_api_key,
+        gemini_api_key=args.gemini_api_key,
+        ollama_base_url=args.ollama_base_url,
+    )
+    if getattr(args, "retriever", None):
+        kwargs["retriever"] = args.retriever
+    if getattr(args, "metadata_filter", None):
+        kwargs["metadata_filter"] = args.metadata_filter
+    return RAG(**kwargs)
+
+
+def _parse_metadata_filter(raw: str | None) -> dict | None:
+    if not raw:
+        return None
+    try:
+        return json.loads(raw)
+    except json.JSONDecodeError as e:
+        raise argparse.ArgumentTypeError(f"Invalid JSON for metadata filter: {e}") from e
+
+
+def _add_provider_args(parser: argparse.ArgumentParser) -> None:
+    parser.add_argument(
+        "--llm-provider",
+        default="openai",
+        choices=["openai", "azure_openai", "anthropic", "gemini", "ollama"],
+        metavar="PROVIDER",
+        help="LLM backend (default: openai). See: rag-python docs providers",
+    )
+    parser.add_argument(
+        "--llm-model",
+        default=None,
+        metavar="MODEL",
+        help="LLM model or Azure deployment name (default: from env LLM_MODEL)",
+    )
+    parser.add_argument(
+        "--embedding-provider",
+        default="openai",
+        choices=["openai", "azure_openai", "ollama", "local"],
+        metavar="PROVIDER",
+        help="Embedding backend (default: openai). Use local for offline embeddings",
+    )
+    parser.add_argument(
+        "--embedding-model",
+        default=None,
+        metavar="MODEL",
+        help="Embedding model name (default: from env EMBEDDING_MODEL)",
+    )
+    parser.add_argument(
+        "--ollama-base-url",
+        default=None,
+        metavar="URL",
+        help="Ollama server URL (default: http://localhost:11434 or OLLAMA_BASE_URL)",
+    )
+    parser.add_argument("--azure-endpoint", default=None, help="Azure OpenAI endpoint URL")
+    parser.add_argument("--azure-api-key", default=None, help="Azure OpenAI API key")
+    parser.add_argument(
+        "--azure-api-version",
+        default=None,
+        help="Azure API version (default: 2023-09-01-preview)",
+    )
+    parser.add_argument("--openai-api-key", default=None, help="OpenAI API key (overrides env)")
+    parser.add_argument("--anthropic-api-key", default=None, help="Anthropic API key")
+    parser.add_argument("--gemini-api-key", default=None, help="Gemini API key")
+
+
+def _add_search_args(parser: argparse.ArgumentParser) -> None:
+    parser.add_argument(
+        "--retriever",
+        choices=["vector", "multi_query", "hybrid"],
+        default=None,
+        metavar="MODE",
+        help=(
+            "Retrieval mode: vector (single query), multi_query (default, with rewriting), "
+            "or hybrid (BM25+vector; requires pip install rag-python[hybrid])"
+        ),
+    )
+    parser.add_argument(
+        "--metadata-filter",
+        type=_parse_metadata_filter,
+        default=None,
+        metavar="JSON",
+        help='Filter chunks by metadata, e.g. \'{"filename": "policy.pdf"}\'',
+    )
+
+
+def _make_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(
+        prog="rag-python",
+        description=(
+            "Production-grade RAG for Python — ingest documents, ask questions, "
+            "get grounded answers with multi-LLM support."
+        ),
+        epilog=CLI_EPILOG,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    parser.add_argument("--version", action="version", version=f"%(prog)s {__version__}")
+    sub = parser.add_subparsers(dest="command", required=True, metavar="COMMAND")
+
+    ing = sub.add_parser(
+        "ingest",
+        help="Load files into the vector store (chunk + embed)",
+        description=(
+            "Ingest one or more files or directories into the ChromaDB vector store.\n"
+            "Supported formats: .txt .md .pdf .docx .csv .json .html"
+        ),
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog=(
+            "examples:\n"
+            "  rag-python ingest ./data --reindex\n"
+            "  rag-python ingest policy.pdf handbook/ --embedding-provider local"
+        ),
+    )
+    ing.add_argument(
+        "paths",
+        nargs="+",
+        metavar="PATH",
+        help="File or directory paths to ingest",
+    )
+    ing.add_argument(
+        "--reindex",
+        action="store_true",
+        help="Delete existing vectors before ingesting (fresh index)",
+    )
+    _add_provider_args(ing)
+
+    q = sub.add_parser(
+        "query",
+        help="Ask a question against ingested documents",
+        description=(
+            "Run the full RAG pipeline: retrieve relevant chunks, generate an answer, "
+            "optionally stream tokens and show sources."
+        ),
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog=(
+            "examples:\n"
+            '  rag-python query "How many days of annual leave?"\n'
+            "  rag-python query \"PTO policy\" --stream -v\n"
+            '  rag-python query "benefits" --retriever hybrid --metadata-filter \'{"filename": "hr.pdf"}\''
+        ),
+    )
+    q.add_argument(
+        "question",
+        nargs="+",
+        metavar="QUESTION",
+        help="Question text (multiple words are joined)",
+    )
+    q.add_argument(
+        "--no-multi-query",
+        action="store_true",
+        help="Use single-query vector retrieval (same as --retriever vector)",
+    )
+    q.add_argument(
+        "--stream",
+        action="store_true",
+        help="Stream answer tokens to stdout as they are generated",
+    )
+    q.add_argument(
+        "-v",
+        "--verbose",
+        action="store_true",
+        help="After the answer, print evaluation scores and top source paths",
+    )
+    _add_provider_args(q)
+    _add_search_args(q)
+
+    docs = sub.add_parser(
+        "docs",
+        help="Show user documentation in the terminal",
+        description="Print built-in help topics. Full docs: https://github.com/RaghavOG/rag-python/tree/main/docs",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="topics: " + ", ".join(list_topics()),
+    )
+    docs.add_argument(
+        "topic",
+        nargs="?",
+        default="quickstart",
+        choices=list_topics(),
+        metavar="TOPIC",
+        help="Documentation topic (default: quickstart)",
+    )
+    docs.add_argument(
+        "--list",
+        action="store_true",
+        help="List all available documentation topics",
+    )
+
+    return parser
+
+
+def main(argv: list[str] | None = None) -> None:
+    parser = _make_parser()
+    args = parser.parse_args(argv)
+
+    if args.command == "docs":
+        if args.list:
+            print_topic_list()
+        else:
+            print_topic(args.topic)
+        return
+
+    if args.command == "ingest":
+        rag = _build_rag(args)
+        n = rag.ingest(args.paths, reindex=args.reindex)
+        print(f"Ingested {n} chunks.")
+        return
+
+    if args.command == "query":
+        rag = _build_rag(args)
+        question = " ".join(args.question)
+        retriever = args.retriever
+        if retriever is None and args.no_multi_query:
+            retriever = "vector"
+        search = replace(
+            rag.config.search,
+            retriever=retriever or rag.config.search.retriever,
+            metadata_filter=args.metadata_filter or rag.config.search.metadata_filter,
+        )
+        if args.stream:
+            stream = rag.query_stream(question, search=search)
+            for token in stream:
+                print(token, end="", flush=True)
+            print()
+            result = stream.result
+            if args.verbose:
+                print("\n--- evaluation ---")
+                print(result.evaluation)
+                print("\n--- sources ---")
+                for s in result.sources[:5]:
+                    print(s.get("metadata", {}).get("source", ""), "score:", s.get("score"))
+            return
+
+        ans = rag.query(question, search=search)
+        print(ans.text)
+        if args.verbose:
+            print("\n--- evaluation ---")
+            print(ans.evaluation)
+            print("\n--- sources ---")
+            for s in ans.sources[:5]:
+                print(s.get("metadata", {}).get("source", ""), "score:", s.get("score"))
+
+
+if __name__ == "__main__":
+    main(sys.argv[1:])