Ubibrowser KeyError

[kkaris]

When running sources.ubibrowser.api.process_from_web, there is a KeyError raised, likely due to updated headers in the latest data from the Ubibrowser.

_Update: the base URL needs to be updated as well. The new URL is http://ubibrowser.bio-it.cn/ubibrowser_v3/Public/download/literature/_

http://ubibrowser.bio-it.cn/ubibrowser_v3/home/download

To re-create the error:

from indra.sources import ubibrowser
up = ubibrowser.process_from_web()

The error output (the Pandas part of the stack trace is omitted for brevity):

KeyError                                  Traceback (most recent call last)
Input In [2], in <cell line: 1>()
----> 1 up = ubibrowser.process_from_web()

File ~/repos/indra/indra/sources/ubibrowser/api.py:23, in process_from_web()
     21 e3_df = pandas.read_csv(E3_URL, sep='\t')
     22 dub_df = pandas.read_csv(DUB_URL, sep='\t')
---> 23 return process_df(e3_df, dub_df)

File ~/repos/indra/indra/sources/ubibrowser/api.py:65, in process_df(e3_df, dub_df)
     49 """Process data frames containing UbiBrowser data.
     50 
     51 Parameters
   (...)
     62     extracted in its statements attribute.
     63 """
     64 up = UbiBrowserProcessor(e3_df, dub_df)
---> 65 up.extract_statements()
     66 return up

File ~/repos/indra/indra/sources/ubibrowser/processor.py:16, in UbiBrowserProcessor.extract_statements(self)
     13 for df, stmt_type in [(self.e3_df, Ubiquitination),
     14                       (self.dub_df, Deubiquitination)]:
     15     for _, row in df.iterrows():
---> 16         stmt = self._process_row(row, stmt_type)
     17         if stmt:
     18             self.statements.append(stmt)

File ~/repos/indra/indra/sources/ubibrowser/processor.py:26, in UbiBrowserProcessor._process_row(row, stmt_type)
     20 @staticmethod
     21 def _process_row(row, stmt_type):
     22     # Note that even in the DUB table the subject of the statement
     23     # is called "E3"
     24     # There are some examples where a complex is implied (e.g., BMI1-RNF2),
     25     # for simplicity we just ignore these
---> 26     if '-' in row['E3AC']:
     27         return None
     28     subj_agent = get_standard_agent(row['E3GENE'], {'UP': row['E3AC']})

[...]

KeyError: 'E3AC'

Inspecting the row variable with debug in IPython reveals the following data structure:

NUMBER                                1
SwissProt ID (E3)            ADO1_ARATH
SwissProt ID (Substrate)    APRR1_ARATH
SwissProt AC (E3)                Q94BT6
SwissProt AC (Substrate)         Q9LKL2
Gene Symbol (E3)                   ADO1
Gene Symbol (Substrate)           APRR1
SOURCE                          MEDLINE
SOURCEID                       22199232
SENTENCE                          E3Net
E3TYPE                            Other
COUNT                                 1
type                              Other
species                      A.thaliana
Name: 0, dtype: object

[bgyori]

It sounds like they renamed their columns and changed E3AC to SwissProt AC (E3) so we would have to update the code accordingly.

[bgyori]

I fixed all the issues on the db-sources-updates branch.

[kkaris]

Resolved on #1423