pybiotech 0.3.0

A collection of reusable python biotech library from AI Lingues.

AI Lingues Biotech Python Library

pybiotech is a biology-focused toolkit from AI Lingues that builds on RDKit, PubChem, and UniProt to deliver molecule optimization, fingerprinting, structured loaders, and request modeling for data-driven biomedical workflows.

---

Key Capabilities

• Molecular engineering with ETKDG embedding, MMFF/UFF optimization, surface and hydrophobic metrics, ring counts, fingerprints, and pharmacophore distances.
• Binary serialization that wraps rdkit.Chem.Mol objects in a CRC32-protected container suitable for IPC, sockets, or shared memory.
• Loader stack covering file, directory, and text SDF parsing plus UniProt XML streaming with error tolerance and namespace handling.
• PubChem request builder that unifies Input, Operation, Output, and Query segments, enforces validation rules, and can reverse-parse existing URLs.
• Schema modeling using Pydantic and xsdata so UniProt query fields and PubChem domain pieces stay type-safe.

---

📦 Installation

Install from PyPI

pip install pybiotech

Requirements

• Python 3.10 or newer (3.11 recommended)
• RDKit for molecule processing
• Additional dependencies declared in pyproject.toml / requirements.txt (xsdata, lxml, pydantic, scipy, numpy, etc.)

---

📚 Quick Start

1. Embed, optimize, and serialize a molecule

from rdkit import Chem
from pybiotech.core.molecule.optimizer import Optimizer

mol = Chem.MolFromSmiles("c1ccccc1O")
mol3d, ok, meta = Optimizer.embed_and_optimize(mol, random_seed=42)
blob = Optimizer.to_serialize([mol3d, None], include_props=True, with_checksum=True)
print("converged:", ok, "method:", meta["method"])

2. Read SDF records from a directory

from pybiotech.loaders.sdf_loader import SDFLoader

mols, expected, actual = SDFLoader.readDataFromDir("data/sdf", recursive=True)
print(f"expected {expected}, loaded {actual} valid molecules")

3. Build a PubChem REST URL

from pybiotech.classes.nih.https.pubchem.input_spec import InputSpec
from pybiotech.classes.nih.https.pubchem.opera_spec import OperationSpec
from pybiotech.classes.nih.https.pubchem.output_spec import OutputSpec
from pybiotech.classes.nih.https.pubchem.query_options import QueryOptions
from pybiotech.classes.nih.https.pubchem.request import PubChemRequest

input_spec = InputSpec(domain="compound", namespace="cid", identifiers="2244")
operation_spec = OperationSpec(operation="property", tags=["MolecularFormula", "InChIKey"], domain="compound")
output_spec = OutputSpec(output_format="JSON")
query_options = QueryOptions(record_type="3d", image_size="large")

req = PubChemRequest(
    input_spec=input_spec,
    operation_spec=operation_spec,
    output_spec=output_spec,
    query_options=query_options,
)
print(req.build_url())

---

Module overview

Core (molecule)

Module	Purpose	Docs
core.molecule.calculator	Geometry, surface, fingerprint, and pharmacophore helpers for Chem.Mol.	docs/core/molecule/calculator.md
core.molecule.optimizer	ETKDG embedding plus MMFF/UFF optimization, serialization container, and checksum-aware deserialization.	docs/core/molecule/optimizer.md

Loaders

Module	Purpose	Docs
loaders.sdf_loader	Stream file/dir/text SDF content with index filtering, error tolerance, and warning control.	docs/pybiotech/loaders/sdf_loader.md
loaders.uniprot_loader	Auto namespace detection and entry-by-entry UniProt XML parsing.	docs/pybiotech/loaders/uniprot_loader.md
loaders.uniprot_query_field_loader	Load QueryField definitions from JSON/text and build validated term:value fragments.	docs/pybiotech/loaders/uniprot_query_field_loader.md

Classes (models and request helpers)

Module	Purpose	Docs
classes.uniprot.https.uniprot.org.uniprot_query_field	Pydantic QueryField model with recursive siblings/items.	docs/classes/uniprot/https/uniprot/org/uniprot_query_field.md
classes.uniprot.https.uniprot.org.uniprot	xsdata-generated dataclasses and enums for UniProt Entry, Protein, Sequence, etc.	docs/classes/uniprot/https/uniprot/org/uniprot.md
classes.nih.https.pubchem.*	Input, Operation, Output, Query, Request, and Identifier/domain modeling for PubChem PUG REST.	docs/classes/nih/https/pubchem

---

📖 Documentation

Detailed API references are under docs/:

• docs/core/molecule/ - optimizer and calculator references
• docs/pybiotech/loaders/ - SDF/UniProt loader documentation
• docs/classes/nih/https/pubchem/ - PubChem URL and query modeling
• docs/classes/uniprot/https/uniprot/org/ - UniProt schema reference

---

🏗️ Project layout

• pybiotech/
  • pybiotech/
    • core/
      • molecule/
    • loaders/
    • classes/
  • docs/
  • scripts/
  • README.md
  • README_Example.md
  • pyproject.toml
  • requirements.txt

---

📝 License & Usage Terms

Package Usage License

The PyBiotech package (the binary package installed via pip) is released under the MIT License:

• ✅ Free Use – May be used freely in personal and enterprise projects
• ✅ Commercial Use – Allowed in commercial products and services
• ✅ Free Distribution – You may freely distribute and redistribute the package
• ✅ No Usage Restrictions – No fees or additional authorization required

Source Code Protection Terms

Important Notice: The source code of this project is private property and protected by intellectual property law:

• ❌ Source Code Not Public – The source code is not publicly available
• ❌ No Reverse Engineering – Decompilation, reverse engineering, or disassembly of the package is strictly prohibited
• ❌ No Source Distribution – You may not obtain, copy, or distribute the source code in any form
• ❌ No Modified Redistribution – You may not modify the package and redistribute it In short: You are free to use our package (including for commercial purposes), but please respect our source code intellectual property rights.

📧 Contact Us

Website: https://www.ailingues.com

Email: [e-mail]

Technical Support: For any questions or suggestions, please contact us via email.

---

Made with ❤️ by AI Lingues Team