Three-letter code extraction doesn't work in ProDy 2.6.1 under Python 3.12

[lokapal]

Hello!

I again try to use prody to extract from mmCIF/PDB three-letter codes. Right now in Python 3.12 and ProDy 2.6.1 installed by pip it simply doesn't work.

The code

from prody import *
polys = parsePDBHeader('1bkv', 'polymers', threeLetter=True)

for prot in polys:
    SEQRES=prot.sequence
    print (SEQRES)

ag = parsePDB('1bkv', compressed=True)
for chain in ag.iterChains():
    print(chain.getSequence(threeLetter=True))
ag.ca.getSequence(threeLetter=True)

produces output:

PPGPPGPPGITGARGLAGPPGPPGPPGPPG
PPGPPGPPGITGARGLAGPPGPPGPPGPPG
GPGPGITGARGLAGPGPGPGPG
PGPGPGITGARGLAGPGPGPGPG
PGPGPGITGARGLAGPGPGPGPG

that is, obviously, wrong

Moreover, I've dived in getSequence source code at
http://www.bahargroup.org/prody/_modules/prody/atomic/chain.html#Chain.getSequence
and I don't see any signs of "threeLetter" keyword processing.

[lokapal]

At least I've captured the problem - threeLetter renamed to the longSeq.
Please, update the docs!

[lokapal]

Still bug present - I cannot mix short and long requests for long sequence - the code

ag = parsePDB(ACC, compressed=True)
for chain in ag.iterChains():
    currid = chain.getChid()
    sequence = chain.getSequence()
    print (f"{ACC}.{currid}\t{sequence}")
    sequence = chain.getSequence(longSeq=True)
    print (f">{ACC}.{currid}\n{sequence}")

always produces short output, the code

ag = parsePDB(ACC, compressed=True)
for chain in ag.iterChains():
    currid = chain.getChid()
    sequence = chain.getSequence(longSeq=True)
    print (f"{ACC}.{currid}\t{sequence}")
    sequence = chain.getSequence(longSeq=False)
    print (f">{ACC}.{currid}\n{sequence}")

always produces long output

[ccwmca]

When you use the getSequence method you set the attribute of the chain instance as the sequence that is calculated (accessed via chain_instance.seq_) in either one-letter or three-letter code, depending on how you have set longSeq. Then, next time you run getSequence it activates this part of the getSequence method:

elif self._seq:
    seq = self._seq

Which just returns the sequence as you have already calculated it. I can see why it could be useful, but ultimately my opinion is that this part of the code is redundant; I have commented it out and have achieved the correct behaviour. Alternatively, set chain_instance._seq = None after using getSequence and then when you call it again it should re-calculate the sequence correctly with the correct longSeq value.