Implemented changes described in comments in the previous PR

[luissian]

The following changes were made for this PR:

• Biopython is used to verify if a nucleotide seq can be converted to protein.
• Error messages while converting are fetched from biopython
• new parameter in analyze schema "cpus" is added
• matching by allele name in prokka gtf output with the one in the gene file
• Crated box plot for variability on allele length
• Bar plot created for graphic showing range of alleles (x axis) and number of genes( y axis)
• Stored in memory the annotation for each gene and save a tar.gz file with all
• Included liting in actions
• Implemented in code the comments requested on the previous PR
• Replaced deprecated setup.py with pyproject.toml
• Update actions to test analyze schema

[saramonzon]

Made some comments for you to check!
Let's see if that test finishes correctly!

[saramonzon]

I feell that maybe this can be done without a tmp_dict?
Something like this?

 value_to_keys = defaultdict(list)

    # Iterate over items and group keys by value
    for key, value in my_dict.items():
        value_to_keys[value].append(key)

    # Filter out values with only one key (no duplicates)
    duplicates = {value: keys for value, keys in value_to_keys.items() if len(keys) > 1}

And the use the duplicates to create the a_quality

[luissian]

yes, code is nicer for getting only which are duplicated, however, I need to create another loop for getting the allele_id based on the sequence value to update the a_quality dictionary for setting the bad quality reason, which requires more CPU time. Because it should be like this:

value_to_keys = defaultdict(list)

# Iterate over items and group keys by value
for key, value in my_dict.items():
    value_to_keys[value].append(key)

# Filter out values with only one key (no duplicates)
duplicates = {value: keys for value, keys in value_to_keys.items() if len(keys) > 1}

# Additional code
for seq_id, seq_value in allele_seq.items():
      if seq_value in duplicates:
          for id in duplicates[seq_value]:
              a_quality[id]["quality"] = "Bad quality"
              a_quality[id]["reason"] = "Duplicate allele"

Please correct me if I misunderstood.

[saramonzon]

mm don't understand, if you keep the seq_id in the duplicates list, you don't need to iterate over that, right?
Maybe you need to twist a little this in order to keep the seq_ids instead of the seq_values:

duplicates = {value: keys for value, keys in value_to_keys.items() if len(keys) > 1}

[luissian]

I was thinking on this code

            value_to_keys = defaultdict(list)
            for rec_id, seq_value in allele_seq.items():
                value_to_keys[seq_value].append(rec_id)
                if len(value_to_keys[seq_value]) > 1:
                    a_quality[rec_id]["quality"] = "Bad quality"
                    a_quality[rec_id]["reason"] = "Duplicate allele"
                    if self.remove_duplicated:
                        bad_quality_record.append(rec_id)

[luissian]

Updated this part of code having only 1 loop and without any temporary dictionary

[saramonzon]

can't you include this check above so we only iterate one time through the dict?
Maybe changing the if clause position?

[luissian]

I can not as I need to remove it, for including in the bad quality record , once I know that it is duplicated

[saramonzon]

I think you can do a double list comprenhension here, but not sure

[luissian]

Yes, I will modify it with this code

        labels = taranis.utils.POSIBLE_BAD_QUALITY
        values = [stats_df[item].sum() for item in labels]

[luissian]

it is implemented this solution in code

[saramonzon]

Suggested change

	create-conda-env:
	tests:

[luissian]

test file updated to use the pyroject.toml file instead of setup.py

- name: Activate env and install taranis
        run: |
          source $CONDA/etc/profile.d/conda.sh
          conda activate taranis_env
          poetry install
          taranis analyze-schema -i  test/MLST_listeria -o analyze_schema_test  --cpus 1 --output-allele-annot --remove-no-cds --remove-duplicated --remove-subset