Skip to content

GitLab

Closed
Created by Oriol Guitart@oguitartOwner

Preprocess scripts

A directory structure is created for the preprocess/signature0 step. One directory for each dataset that contains a run.py file and maybe more files.

@mbertoni There is a final step in all preprocesses scripts. This final step has three tasks and the input are two maps(an inchikey to inchi map and the inchikey to raw map):

  1. Find from the inchikeys in the map which are not in structure table and add them to the table
  2. For each of the missing inchikeys draw its image and store it in IN/CH/INCHIKEY/
  3. Store the raws in the sign0.h5

The first task requires some complex query to the DB, we should decide if we use our ORM framework or a psycopg query. We also need to find a place to put common methods like drawing.