[BioPython] problem with GenBank.NCBIDictionary ?

Jeffrey Chang jchang@smi.stanford.edu
Tue, 21 May 2002 00:02:37 -0700


On Thu, May 16, 2002 at 10:12:14AM +1000, Cath Lawrence wrote:
> I actually got several records with partial HTML markup from NCBI, both 
> when downloaded through biopython and direct from the website. So it 
> looks like it is entrez at fault rather than biopython. They were all 
> anchor tags, mostly in the feature areas. I ended up removing them 
> manually with sed.

Yeah, I think I've seen this before.  I'm glad someone else noticed it
too -- I thought I was losing my mind!  Do you think it would help if,
whenever we get back HTML'd data, we try the query again a few more
times, to see if the HTML goes away?  Or, would it be better to just
rip out the HTML tags and make sure the results "look like" a genbank
record, to make sure Entrez is not telling me that there's no such
record?

Jeff