Interactive hub demo

Run the full OpenAlgo flow on a seeded Hub paper

This walkthrough starts from the Hub example for Random Forest Models for ADMET Property Prediction with RDKit Descriptors and follows the same review, gap-analysis, generation, and validation sequence as a real translation session.

hub demoneeds_review

Descriptor classification walkthrough

Review the extracted method, inspect the reproducibility gaps, and generate a template-first Python project for a descriptor-based ADMET classifier.

Template family

fingerprint_descriptor_classification

Detected domain

supported

Confidence

91/100

Primary metric

ROC-AUC

Open source gap reportView template coverage

This walkthrough runs entirely in the browser. It does not create a saved translation session or require Supabase.

01pending

Review extraction

Inspect the seeded method record and edit any field that feels too optimistic.

02pending

Generate template

Produce the scaffold, config, reproducibility gap report, and validation summary.

03pending

Inspect outputs

Browse the generated files exactly where a real translation session would expose them.

Reproducibility gap report

The gap report updates live as you edit the extraction.

12 Specified
0 Partial
1 Missing

Specified

Dataset

Curated ADMET assay panel

Data

Molecular features

RDKit molecular descriptors

Methods

Model

Classical ML (Random Forest, SVM, XGBoost, etc.)

Methods

Split strategy

Repeated random 80/20 split

Validation

Primary metric

ROC-AUC

Results

Random seed

Random seed mentioned in the paper

Salt stripping

Salt handling discussed in the paper

Hyperparameter tuning

Hyperparameter search strategy described

RDKit version

Exact RDKit version specified

Feature scaling

Feature scaling/normalization mentioned

Class imbalance handling

Class imbalance strategy described

Train/test ratio

Exact split ratio reported

Partially specified

No items in this column.

Not specified

Stereoisomer handling

Not addressed — keeping stereochemistry as-is

Demo status

Demo extraction is ready for review. Confirm the fields, then generate the template artifact set.

Save review before generation so the flow mirrors the real translation state machine.

Reviewer note

Notes stay local to this demo and are not sent anywhere.

Random Forest Models for ADMET Property Prediction with RDKit Descriptors demo | OpenAlgo | OpenAlgo