kexinhuang12345 / espf Goto Github PK

The prune_stats method considers the case when freq is less than 0. The only spot where frequency is reduced is in update_pair_statistics where frequency of neighboring pairs is reduced because of the merging.

However why could it be reduced to less than 0 given that changes produced by replace_pair reflects exactly which co-occurance of the merged pair is located in the original context?

Hi, I would like to know if there are any special considerations for marking the last character with </w> since this will affect statistics of related pairs. Is it chemically inspired?

An example being C C O<\w> won't match C O by the given regex pattern r'(?<!\S)' + re.escape(first + ' ' + second) + r'(?!\S)'

Hi! I'm trying to generate my own ESPF fingerprint, hence I'm using the notebook generate_drug.ipynb in this repo.
Problem is that the text file chembl_seq.txt is missing. Can you upload it, please?

Hey @kexinhuang12345 hope you will be doing well I have a question below can you please guide me.
I want to create a substructure for DrugBank DB only what should be the value of frequency can you please guide me a little.
Best Regards Waqar Kaleem Khan