deepmedchem 0.3.0b4

Enamine REAL now defaults to the filtered 2026-09-06.2 release in production: approximately 93.41B source reagent combinations, with a 95% interval of 93.19–93.63B. These are estimated combinations, not unique molecular graphs.

• Exact SMILES/SMARTS substructure search is enabled, with all 326 route partitions indexed and final products verified against the original query.
• Sampling, similarity, selections, runs, and substructure share the historical topology policy and exact Ro5/Veber product limits. Price estimates remain available.
• The CLI marks estimated counts with ~ and labels result summaries as source combinations.
• README, live examples, and notebook documentation reflect production. Substructure examples allow an HTTP timeout margin above the server budget, and documentation links point to the current guide.

Install or upgrade:

python -m pip install --upgrade --pre deepmedchem==0.3.0b4

The existing database="enamine" alias and Python search methods continue to work. Complex substructure queries remain bounded by the requested timeout and can return partial results.

Validation: production acceptance checked all 300 returned products independently, with eight positive substructure cases and a bounded empty-result case. CLI/export/documented examples passed. Core, processing, backend and SDK suites passed (335 tests total), along with the SDK/backend contract and documentation build. SDK CI passed on Python 3.9–3.13.

deepmedchem 0.3.0b3

Update the XtalPi VAST vendor contact used by ordering (dmc order / prepare_vendor_requests) and the README database tables from contact@xtalpi.com to VAST@XtalPi.com. No API or behavior changes otherwise. PR #27.

deepmedchem 0.3.0b2

Adds short database names across the Python SDK and CLI, a compact catalog in the documented order, and a detailed catalog with full database IDs, BioSolveIT mappings, availability, and vendor links.

• Accept names such as enamine, freedom, and vast in synchronous and asynchronous requests, selections, and order CSV input.
• Refresh README chemical-space tables, migration guidance, links, and production Enamine examples with restored price estimates.
• Clarify price-estimate assumptions and update VAST H2 2026 lead-time and synthesis-success metadata.
• Replace the blanket CLI delivery estimate with database-specific guidance.

Validation: 88 tests, lint, wheel/sdist builds, Twine metadata checks, and installation smoke tests. CI covers Python 3.9–3.13.

Install or upgrade:

python -m pip install --upgrade deepmedchem==0.3.0b2

deepmedchem 0.3.0b1

Adds normalized molecule-selection support for exact assembled-product predicted-property ranges, separate factorized CP16 and exact OpenADMET values/model identities, and bounded unique-product result telemetry. Includes runnable Lipinski, soft ADMET acquisition, and hard hERG range examples. Removes the obsolete shortlist-multiplier workaround from the public client and examples.

deepmedchem 0.2.0b7

Exact RDKit property filters and experimental predicted-property (CP16) acquisition, through one Selection document.

selection = (
    Selection.from_database("enamine-real-v5a")
    .reference("query", smiles="CCOc1ccc(C(=O)N2CCN(C)CC2)cc1")
    .maximize_similarity("rdkit.ecfp4_tanimoto", reference="query")
    .require_preset("lipinski-ro5/v1")
    .where("rdkit.mol_wt", lte=450, units="Da")
    .acquire_predicted_property("openadmet-herg-pchembl", direction="minimize", keep_fraction=0.25)
    .include("properties", "objective_components")
    .limit(100)
)
result = client.selections.create(selection)

• Selection.acquire_predicted_property(...) with typed per-hit (predicted_value, applicable) and response-level (endpoint_id, model_version, direction, units, qualification, before/after counts) acquisition models
• exact assembled-product RDKit values on hits; the server recalculates and enforces every threshold literally
• the same selection contract covers RDKit-filtered random sampling
• compact CLI result tables, database sizes, and vendor price availability in dmc databases
• README, documented examples and the Agent Skill reference updated

CP16 values are predicted, experimental-acquisition-only ranking signals. They may reduce a similarity shortlist; they are not assay results and never establish that an ADMET threshold is met.

Available on Enamine REAL v5a and Freedom Space 5; read the authenticated catalog for each database's properties, presets and endpoints.

deepmedchem 0.2.0b5

• dmc login now verifies the saved key against the API and warns when it is rejected, instead of reporting success for an expired or revoked key.

pip install --pre deepmedchem==0.2.0b5

deepmedchem 0.2.0b4

Documentation-only release: the README no longer describes how vendor price estimates are derived. Prices remain available on every hit as before. See v0.2.0b3 for the dmc command line, usage, and export features.

pip install --pre deepmedchem==0.2.0b4

deepmedchem 0.2.0b3

Command line

• New short dmc command (the deepmedchem entry point remains as an alias).
• dmc databases lists every searchable space with typical delivery time and the vendor address for orders and quotes.
• dmc usage shows the account plan and today's CHEESE Credit balance, including any active promotion.
• dmc search, dmc substructure, and dmc sample print rank, score, price, product id, and SMILES, and save results with -o file.csv|.sdf|.smi|.json.
• dmc order results.csv prepares price-free vendor quote or order email drafts.

SDK

• Client.usage(), AsyncClient.usage(), and deepmedchem.usage() return a typed Usage model served by the account service (account_url per profile, DEEPMEDCHEM_ACCOUNT_URL override).
• SearchResult.to_csv(), to_sdf() (needs RDKit, pip install "deepmedchem[sdf]"), and to_file().

pip install --pre deepmedchem==0.2.0b3

deepmedchem 0.2.0b2

Production API convergence beta: defaults to https://api.deepmedchem.com, supports the full API v2 surface, shared CHEESE API keys/credits, Python 3.9+, VAST 2026 H2, and documented synchronous/durable execution limits.

deepmedchem 0.2.0b1

First public beta of the official DeepMedChem Python SDK.

Highlights:

• synchronous and asynchronous clients for the hosted DeepMedChem API
• similarity, shape, ESP, substructure, sampling, selection, and durable-run operations
• typed results with whole-dollar pricing
• OS-keyring authentication and application attribution
• Python 3.10–3.13 support

Install:

pip install deepmedchem==0.2.0b1

This is a beta release; APIs may change before 0.2.0.