ragger-simple 0.1.1

Simple vector database operations with Qdrant

Vector Search

A simple Python package for vector database operations using Qdrant.

Features

• Initialize connection with Qdrant vector database
• Parse, chunk, and process text into vector embeddings
• Search for relevant text chunks based on semantic similarity

Installation

pip install vector-search

Usage

CLI Commands

You can use the vector-search console script with three commands: init, process, and search.

init

Initialize the vector database:

vector-search init \
  --collection COLLECTION_NAME \
  [--model MODEL_NAME] \
  [--qdrant-url QDRANT_URL] \
  [--qdrant-key QDRANT_API_KEY] \
  [--qdrant-path QDRANT_PATH]

Options:

• --collection (str, default: "documents")
• --model (str, default: "all-MiniLM-L6-v2")
• --qdrant-url (str) — Qdrant Cloud URL
• --qdrant-key (str) — Qdrant API key
• --qdrant-path (str) — Path for local Qdrant

process

Process a JSON file of documents and add to the database:

vector-search process \
  --input INPUT_FILE.json \
  [--collection COLLECTION_NAME] \
  [--chunk-size CHUNK_SIZE] \
  [--overlap OVERLAP] \
  [--qdrant-url QDRANT_URL] \
  [--qdrant-key QDRANT_API_KEY] \
  [--qdrant-path QDRANT_PATH]

Options:

• --input (str, required) — Path to JSON { name: text }
• --collection (str, default: "documents")
• --chunk-size (int, default: 200) — words per chunk
• --overlap (int, default: 50) — overlapping words
• --qdrant-url, --qdrant-key, --qdrant-path — same as init

search

Search for relevant chunks:

vector-search search \
  --query QUERY_TEXT \
  [--collection COLLECTION_NAME] \
  [--k K] \
  [--output OUTPUT_FILE.json] \
  [--qdrant-url QDRANT_URL] \
  [--qdrant-key QDRANT_API_KEY] \
  [--qdrant-path QDRANT_PATH]

Options:

• --query (str, required) — query text
• --collection (str, default: "documents")
• --k (int, default: 5) — number of results
• --output (str) — path to write JSON output (defaults to stdout)
• --qdrant-url, --qdrant-key, --qdrant-path — same as init

Python API

Import and initialize VectorDB:

from ragger_simple.db import VectorDB

db = VectorDB(
    collection_name="my_documents",
    model_name="all-MiniLM-L6-v2",
    qdrant_url=None,
    qdrant_api_key=None,
    qdrant_path=None,
    qdrant_timeout=500.0,
)

Constructor parameters:

• collection_name (str) — Qdrant collection name
• model_name (str) — sentence-transformers model
• qdrant_url (str, optional) — cloud URL
• qdrant_api_key (str, optional) — cloud API key
• qdrant_path (str, optional) — local path
• qdrant_timeout (float, default: 500) — request timeout

Methods:

db.add_documents(
    documents: Dict[str, str],
    chunk_size: int = 200,
    overlap: int = 50,
)

• documents — dict mapping doc names to text
• chunk_size — words per chunk
• overlap — overlapping words

results = db.search(
    query: str,
    k: int = 5,
) -> List[Dict]

• query — query text
• k — number of results

Example:

documents = {
    "Article 1": "This is the content of article 1...",
    "Article 2": "This is the content of article 2..."
}
db.add_documents(documents, chunk_size=200, overlap=50)
results = db.search("your query here", k=5)
print(results)

License

MIT