access to data

Access to data

The CC resource provides a very simple API to query the data within the resource. In brief, the API facilitates the following:

Given one or more inchikeys:
It lists the corresponding datasets.
It returns the InChI structure(s).
If a dataset is specified, it returns the signatures (sign*) or vectorial data (neig*, clus*, etc.) of choice.
Given one or more datasets:
It lists the corresponding inchikeys.
It returns the signatures (sign*) or vectorial data (neig*, clus*, etc.) of choice.
It deletes the dataset from the resource (might be convenient when we are testing datasets).
Given one or more levels or coordinates:
It lists the corresponding datasets.
It indicates the exemplar dataset(s).

Eventually, this package should be able to perform predict() methods, but this will not happen in the near future.

To install the chemicalchecker package, simply do:

pip install --index http://devpi.sbnb.org:3141/root/prod/ chemicalchecker

⚠ I haven't thought about this part too much, yet. So this is a very preliminary proposal. It is just to get an idea.

Let's get started by simply fetching 2D chemical information for a molecule of interest.

import chemicalchecker as cc

ik = "RZVAJINKPMORJF-UHFFFAOYSA-N"

cc.fetch(inchikey = ik, dataset = "A1.001", type = "sign1")

To directly extract the exemplary signature, we may simply do:

cc.fetch(inchikey = ik, dataset = "A1", type = "sign1")

Now let's focus on a certain dataset of interest. Say, for instance, that we want signatures of type 2.

ds = "B1.001"

cc.fetch(dataset = ds, type = "sign2")