indian-address-parser 0.1.1

Parse raw Indian address strings into structured fields using a fine-tuned Qwen3 LoRA adapter

indian-address-parser

Parse raw, unstructured Indian address strings into 13 structured fields using a Qwen3-0.6B model fine-tuned with LoRA. Model weights are downloaded automatically from Hugging Face — this package ships only inference code, no weights.

Input:  "FLAT NO.32, UTTARA TOWERS, MG ROAD GUWAHATI , Kamrup Unclassified AS 781029"
Output: {"houseNumber": "FLAT NO.32", "houseName": "UTTARA TOWERS", "poi": null,
         "street": "MG ROAD", "subsubLocality": null, "subLocality": null, "locality": null,
         "village": null, "subDistrict": null, "district": "Kamrup", "city": "GUWAHATI",
         "state": "AS", "pincode": "781029"}

Install

pip install indian-address-parser

Usage

Python

from indian_address_parser import AddressParser

parser = AddressParser()  # downloads model weights from HF on first use
result = parser.parse("FLAT NO.32, UTTARA TOWERS, MG ROAD GUWAHATI , Kamrup Unclassified AS 781029")
print(result)

# Batch
results = parser.parse_batch([addr1, addr2, addr3])

CLI

# Single address
indian-address-parser "FLAT NO.32, UTTARA TOWERS, MG ROAD GUWAHATI , Kamrup Unclassified AS 781029"

# Batch from stdin
cat addresses.txt | indian-address-parser --stdin

# Batch from a file, JSONL output
indian-address-parser --file addresses.txt --out results.jsonl

Fields

houseNumber, houseName, poi, street, subsubLocality, subLocality,
locality, village, subDistrict, district, city, state, pincode

Any field not present in the address is null. If the model output can't be parsed as JSON, all fields are null and a _parse_error key holds the raw model output.

Model details, evaluation metrics, and known limitations

See the model card for training data, LoRA config, per-field evaluation results (100% JSON parse rate, 82.4% mean field accuracy on held-out test data), and known limitations (locality/subLocality/ subsubLocality/village field-boundary ambiguity, etc.).

Apple Silicon (MLX) users

This package uses transformers+peft, which works on CUDA, MPS, and CPU but is not the fastest path on Apple Silicon. For MLX-native inference, see the mlx/ subfolder of the Hugging Face repo instead.

License

Apache 2.0 (matching the base model, Qwen/Qwen3-0.6B).