Releases: lozuponelab/KEGG_parser
v0.0.12
v0.0.11
KEGG moved the KO reaction from DBLINKS: RN to REACTION, so parser had to be updated to address this. Now, the reactions for a KO are returned as a list of tuples, like the pathways, formatted as (<RN>, <reaction name>) and are NOT nested within DBLINKS.
More detail/an example can be seen in AMON issue lozuponelab/AMON#19 , which this partially addresses.
v0.0.10
Bump version for PyPI
v0.0.9
Addresses bugs from updated KEGG fields
v0.0.8
KEGG_parser now has an option to skip attempting the asynchronous download, which has been causing issues as KEGG changes their request rate limits. Also added new fields to the "not captured".
v0.0.7
Fix PyPI build issues
v0.0.6
Sometimes, when downloading information for large numbers of KOs from the KEGG API, the API will forbid downloads because there are too many requests (see this issue in AMON). This isn't an ideal fix, but with these updates, if the asynchronous requests are forbidden, the KOs will be downloaded sequentially instead of asynchronously. This is quite a bit slower, but it gets around the issue.
The request rate limit of the API isn't clearly laid out, so this is what we have for now. Perhaps a future approach should have some batch size limit for asynchronous requests.
0.0.5
Fixes PyPI name incompatibility via version update
0.0.4
Actual twine-compatible build for PyPI publishing
0.0.3
Sets up twine-compatible CI