qmem 0.0.5

Qdrant wrapper for simple ingestion & retrieval with pluggable embeddings

qmem

qmem for easy ingestion and retrieval with embeddings using qdrant Supports both CLI and Python API.

---

🚀 Installation

pip install -e .

---

⚙️ CLI Usage

1. Init (configure keys & embedding model)

qmem init

2. Ingest data

qmem ingest

You’ll be prompted for:

• collection_name
• data file path (JSON or JSONL)
• field to embed (e.g. query, response, sql_query, doc_id)
• payload fields (comma-separated, leave empty to keep all)

3. Retrieve results

qmem retrieve

You’ll be prompted for:

• collection_name
• query
• top_k (number of results to return)

---

🐍 Python API

import qmem as qm

# Create a collection
qm.create(collection_name="test_learn", dim=1536, distance_metric="cosine")

# Ingest data from a file
qm.ingest(
    file="/home/User/data.jsonl",
    embed_field="sql_query",
    payload_field="query,response",  # optional, keep these fields in payload
)

# Retrieve results (pretty table by default)
table = qm.retrieve(query="list customers", top_k=5)
print(table)

---

📄 License

MIT