cvua-rrw / taxidtools Goto Github PK

The information present in the delnodes.dmp and merged.dmp files should be (optionaly?) leveraged by the ncbi constructor to:

• redirect old taxid to the new Node
• raise an informative error when trying to access a deleted Node

Test cases:
2923454 should point to 89248 (Equus quagga)
Any taxid in delnodes should raise anerror

This could be implemented by introducing new BaseNode subclasses.
The Taxonomy __getitem__ method could then check the BaseNode subclass and return the correct object (redirected to another Node for merged nodes or raise anerror for deleted nodes)

the Node._children argument is deleted during json creation.
Will be fixed in the next version.

Workaround meanwhile is to reload the json file that was just created.

Providing each path is an unescessatry chore.It should be possible tojust use the tar archive with the tarfile library

Provided that names are unique, provide a method that returns the Taxid for a given node name.

May need to implement a reversing of the dictionnary for that.

First, thank you for this great toolkit, it's super convenient.

I think I might be dealing with a bug.
Version used: 3.1.1

Steps to reproduce:

1. download taxdump
2. Run code:

import taxidTools as txd

TAXDUMP = 'path_to_taxdump/'

tax = txd.read_taxdump(
    f'{TAXDUMP}nodes.dmp',
    f'{TAXDUMP}rankedlineage.dmp',
    f'{TAXDUMP}merged.dmp'
)

# Prune to only keep embryophyta to avoid ambiguity.
tax_pruned = tax.copy() 
tax_pruned.prune(3193)

# Get ancestry of "Liliaceae" from pruned taxonomy. Throws an error. 
tax_id_pruned = tax_pruned.getTaxid('Liliaceae')
tax_pruned.getAncestry(tax_id_pruned)  # Throws an error.

# Get ancestry of "Liliaceae" from original taxonomy. Works fine.
tax_id = tax.getTaxid('Liliaceae')
tax.getAncestry(tax_id)   # Works fine.

#  Both taxonomy ids are the same.
tax_id == tax_id_pruned

Reason is that a taxid argument is automatically given for DummyNode and another one is passed from the json reader.