Load figure caption from the CORD-19 data and add links to PMC. #16
Conversation
…ntly the links only work with PMC new url format, some papers are still using the old url format.
There was a problem hiding this comment.
I really like the to_datetime part of the example as that touches on the new functionality and is simple to understand. I wonder if the rest of the load_csv changes introduce too much complexity to be useful as an example, and that what you're really doing is the start of a research project that uses the pipeline (possibly as a sub-module?) but should be in its own repo?
Also, note the comment about fields vs list_fields; we ought to be able to have one list and just do the right thing based on data type.
Project/pipeline_views.py
Outdated
| field: _nicestr(paper[field]) | ||
| if field in paper["field_order"] | ||
| else paper[field] | ||
| for d in ["field_order", "list_field_order"] |
There was a problem hiding this comment.
Why do field_order and list_field_order need to be separate? I could imagine one might want things with or without lists in any order, not all the non-lists before all the lists.
There was a problem hiding this comment.
i.e. can't we figure this out based on data type?
examples/load_csv.py
Outdated
| label = caption.split(":")[0] | ||
| fignum = int(re.findall("\d+", label)[0]) | ||
| except: | ||
| fignum = int(k.lstrip("TABREF")) + 1 |
There was a problem hiding this comment.
I don't think lstrip is what you're looking for here; consider:
>>> 'TARANTULA'.lstrip('TABREF')
'NTULA'
| @@ -110,12 +110,18 @@ def _nicestr(item): | |||
| if any("," in thing for thing in item): | |||
| joiner = "; " | |||
| return joiner.join(thing for thing in item) | |||
| elif isinstance(item, datetime.datetime): | |||
Currently the links only work with PMC new url format, some papers are still using the old url format.