Python wrapper for the IBM RXN for Chemistry API

A python wrapper to access the API of the IBM RXN for Chemistry website.

Install

From PYPI:

pip install rxn4chemistry

Or directly from the repo:

pip install git+https://github.com/rxn4chemistry/rxn4chemistry.git

Usage

By default, the wrapper connects to the https://rxn.res.ibm.com server. This can be overriden by setting an environment variable. To set a different url, simply do:

export RXN4CHEMISTRY_BASE_URL="https://some.other.rxn.server"

The base url can be directly set when instantiating the RXN4ChemistryWrapper (this will overwrite the environment variable):

api_key = 'API_KEY'
from rxn4chemistry import RXN4ChemistryWrapper

rxn4chemistry_wrapper = RXN4ChemistryWrapper(api_key=api_key, base_url='https://some.other.rxn.server')
# or set it afterwards
# rxn4chemistry_wrapper = RXN4ChemistryWrapper(api_key=api_key)
# rxn4chemistry_wrapper.set_base_url('https://some.other.rxn.server')

Create a project

Get your API key from here and build the wrapper:

api_key = 'API_KEY'
from rxn4chemistry import RXN4ChemistryWrapper

rxn4chemistry_wrapper = RXN4ChemistryWrapper(api_key=api_key)
# NOTE: you can create a project or set an esiting one using:
# rxn4chemistry_wrapper.set_project('PROJECT_ID')
rxn4chemistry_wrapper.create_project('test_wrapper')
print(rxn4chemistry_wrapper.project_id)

Reaction outcome prediction

Running a reaction outcome prediction is as simple as:

response = rxn4chemistry_wrapper.predict_reaction(
    'BrBr.c1ccc2cc3ccccc3cc2c1'
)
results = rxn4chemistry_wrapper.get_predict_reaction_results(
    response['prediction_id']
)
print(results['response']['payload']['attempts'][0]['smiles'])

Extracting actions from a paragraph describing a recipe

Extract the actions from a recipe:

results = rxn4chemistry_wrapper.paragraph_to_actions(
    'To a stirred solution of '
    '7-(difluoromethylsulfonyl)-4-fluoro-indan-1-one (110 mg, '
    '0.42 mmol) in methanol (4 mL) was added sodium borohydride '
    '(24 mg, 0.62 mmol). The reaction mixture was stirred at '
    'ambient temperature for 1 hour.'
)
print(results['actions'])

Retrosynthesis prediction

Predict a retrosynthetic pathway given a product:

response = rxn4chemistry_wrapper.predict_automatic_retrosynthesis(
    'Brc1c2ccccc2c(Br)c2ccccc12'
)
results = rxn4chemistry_wrapper.get_predict_automatic_retrosynthesis_results(
    response['prediction_id']
)
print(results['status'])
# NOTE: upon 'SUCCESS' you can inspect the predicted retrosynthetic paths.
print(results['retrosynthetic_paths'][0])

See here for a more comprehensive example.

Biocatalysed retrosynthesis prediction

Predict a biocatalysed retrosynthetic pathway given a product by specifying the model trained on biocatalysed reactions:

response = rxn4chemistry_wrapper.predict_automatic_retrosynthesis(
    'OC1C(O)C=C(Br)C=C1', ai_model='enzymatic-2021-04-16'
)
results = rxn4chemistry_wrapper.get_predict_automatic_retrosynthesis_results(
    response['prediction_id']
)
print(results['status'])
# NOTE: upon 'SUCCESS' you can inspect the predicted retrosynthetic paths.
print(results['retrosynthetic_paths'][0])

Prediction of reaction properties (atom-to-atom mapping, reaction yield, ...)

Prediction of atom-to-atom mapping (see paper):

response = rxn4chemistry_wrapper.predict_reaction_properties(
    reactions=[
        "CC(C)S.CN(C)C=O.Fc1cccnc1F.O=C([O-])[O-].[K+].[K+]>>CC(C)Sc1ncccc1F",
        "C1COCCO1.CC(C)(C)OC(=O)CONC(=O)NCc1cccc2ccccc12.Cl>>O=C(O)CONC(=O)NCc1cccc2ccccc12",
        "C=CCN=C=S.CNCc1ccc(C#N)cc1.NNC(=O)c1cn2c(n1)CCCC2>>C=CCN1C(C2=CN3CCCCC3=N2)=NN=C1N(C)CC1=CC=C(C#N)C=C1",
    ],
    ai_model="atom-mapping-2020",
)
for predicted_mapping_dict in response["response"]["payload"]["content"]:
    print(predicted_mapping_dict["value"])

Prediction of reaction yields (see paper):

response = rxn4chemistry_wrapper.predict_reaction_properties(
    reactions=[
        "Clc1ccccn1.Cc1ccc(N)cc1.O=S(=O)(O[Pd]1c2ccccc2-c2ccccc2N~1)C(F)(F)F.COc1ccc(OC)c(P([C@]23C[C@H]4C[C@H](C[C@H](C4)C2)C3)[C@]23C[C@H]4C[C@H](C[C@H](C4)C2)C3)c1-c1c(C(C)C)cc(C(C)C)cc1C(C)C.CCN=P(N=P(N(C)C)(N(C)C)N(C)C)(N(C)C)N(C)C.Cc1cc(C)on1>>Cc1ccc(Nc2ccccn2)cc1",
        "Brc1ccccn1.Cc1ccc(N)cc1.O=S(=O)(O[Pd]1c2ccccc2-c2ccccc2N~1)C(F)(F)F.COc1ccc(OC)c(P([C@]23C[C@H]4C[C@H](C[C@H](C4)C2)C3)[C@]23C[C@H]4C[C@H](C[C@H](C4)C2)C3)c1-c1c(C(C)C)cc(C(C)C)cc1C(C)C.CCN=P(N=P(N(C)C)(N(C)C)N(C)C)(N(C)C)N(C)C.COC(=O)c1ccno1>>Cc1ccc(Nc2ccccn2)cc1",
    ],
    ai_model="yield-2020-08-10",
)
for predicted_yield_dict in response["response"]["payload"]["content"]:
    print(predicted_yield_dict["value"])

Create a synthesis and start it on the robot (or simulator)

Create a synthesis from a retrosynthesis sequence:

# Each retrosynthetic path predicted has a unique sequence_id that can
# be used to create a new synthesis
response = rxn4chemistry_wrapper.create_synthesis_from_sequence(
    sequence_id=results['retrosynthetic_paths'][0]['sequenceId']
)
print(response['synthesis_id'])

# get the entire list of actions for the entire synthesis, as well as a tree representation
synthesis_tree, ordered_tree_nodes, ordered_list_of_actions = rxn4chemistry_wrapper.get_synthesis_plan(
    synthesis_id=response['synthesis_id']
)
for action in ordered_list_of_actions:
    print(action)

synthesis_status_result = rxn4chemistry_wrapper.start_synthesis(
    synthesis_id=response['synthesis_id']
)
print(synthesis_status_result['status'])

synthesis_status_result = rxn4chemistry_wrapper.get_synthesis_status(
    synthesis_id=response['synthesis_id']
)
print(synthesis_status_result['status'])

Forward prediction in batch

It is possible to run a batch of forward reaction predictions without linking them to a project:

response = rxn4chemistry_wrapper.predict_reaction_batch(precursors_list=['BrBr.c1ccc2cc3ccccc3cc2c1', 'Cl.c1ccc2cc3ccccc3cc2c1']*5)
# wait for the predictions to complete
time.sleep(2)
print(rxn4chemistry_wrapper.get_predict_reaction_batch_results(response["task_id"]))

NOTE: the results for batch prediction are not stored permanently in our databases, so we strongly recommend to save them since they will expire.

Prediction of multiple reaction outcomes (in batch)

It is also possible to predict multiple forward reaction prediction outcomes in batch:

response = rxn4chemistry_wrapper.predict_reaction_batch_topn(
    precursors_lists=[
        ["BrBr", "c1ccc2cc3ccccc3cc2c1"],
        ["BrBr", "c1ccc2cc3ccccc3cc2c1CCO"],
    ],
    topn=3,
)
# wait for the predictions to complete
time.sleep(2)
print(rxn4chemistry_wrapper.get_predict_reaction_batch_topn_results(response["task_id"]))

NOTE: the results for batch prediction are not stored permanently in our databases, so we strongly recommend to save them since they will expire.

Enable logging

Logging by the library is disabled by default as it may interfere with programmatic uses.

In the very top of the rxn4chemistry_tour.ipynb example notebook you can see a line that enables all logging in the notebook.

import logging
logging.basicConfig(level=logging.INFO, format='%(levelname)s : %(message)s')

This may also enable logging from other libraries. If you wish to selectively enable the logs from rxn4chemistry, consider something like this:

import logging
logger = logging.getLogger("rxn4chemistry")
handler = logging.StreamHandler()
handler.setFormatter(logging.Formatter('%(levelname)s : %(message)s'))
logger.addHandler(handler)
logger.setLevel(logging.DEBUG)