routehunter 1.0.0

Installation

pip install routehunter

Requires Python 3.12 or later.

Quickstart

RouteHunter’s underlying dataset is hosted on Hugging Face. Download it once, then build the app from it:

from routehunter.utils import download_app_data
from routehunter import RouteHunterApp

app_data_dir = download_app_data(to=".")
app = RouteHunterApp.from_data_dir(app_data_dir)

print(app.review())

result = app.search("C#CCOC1=C(C=C(C(=C1)N2C(=O)N3CCCCC3=N2)Cl)Cl")

print(result.found)          # True if a paper or a CASP tool already covers this molecule
print(result.paper_message)  # e.g. "Found 1 paper(s) reporting a route for this molecule"
print(result.paper_report)   # pandas.DataFrame: journal, title, year, doi
print(result.tool_message)   # e.g. "Found 1 tool(s) predicted routes for this molecule"
print(result.tool_report)    # pandas.DataFrame: tool, result, route

result = app.predict("CC(C)Cc1ccc(cc1)C(C)C(=O)O")
print(result.to_dataframe())  # predicted solvability probability per CASP tool, with a link to each tool

papers = app.monitor(year_min=2023, year_max=2025)
print(papers)  # pandas.DataFrame, ranked by predicted probability of reporting a synthesis route

CASP tools currently covered

• AiZynthFinder

• SynPlanner

Project layout

routehunter/        the published package (pip install routehunter) -- everything above lives here
routehunter_build/  internal tooling for building and maintaining the dataset (not published to PyPI)
streamlit/          the Streamlit app behind the live demo

Contributing

RouteHunter is a static, curated dataset – there’s currently no way to add or edit data files directly. All data is updated manually by an administrator after the submitted data has been validated.

Contributions of any kind are welcome:

• Scientists: propose new papers and their target molecules for the dataset, or flag targets with no known route that you’d like tested against open-source CASP tools.

• Developers: propose having your CASP tool integrated into RouteHunter, for use in route search and solvability prediction.

Send contributions to dvzankov@gmail.com, and please include the name you’d like registered as the contributor on your records.

License

MIT – see LICENSE.