search

gittaylor / athenapy

A toolkit for developing and deploying python applications for the UC ATHENA project
Apache License 2.0
1 stars 0 forks source link

Allowing convert.py to use alternate index fields #2

Open rkelly opened 11 years ago

rkelly commented 11 years ago

Currently, convert.py requires that the conversion file have three fields (MRN, CompletedDate, and ApptDate) (lines 117-119). And I believe it assumes that the MRN field will be the index.

However, there are other fields that can be used as index fields. The most likely for CDCG purposes is the AthenaID field.

UCI also has a couple other fields that can be used as indexes (ScrambledId or PatientLocalIdentifier)

Can we make it so "MRN" isnt hardcoded into the convert.py program? If convert.py needs to know what to use as an index, maybe we can have that field be preceded by a special character in the conversion file (like a #).

gittaylor commented 11 years ago

We will have to have at least 1 function that is unique to SFMR, since there are requirements for the SFMR implementation. We should create another function that mimics convertToSFMR() for other conversions. Simply copy the code to another function (e.g. convertToCDCG)...

On Tue, Feb 5, 2013 at 9:13 PM, rkelly notifications@github.com wrote:

Currently, convert.py required that the conversion file have three fields (MRN, CompletedDate, and ApptDate) (lines 117-119).

However, there are other fields that can be used as index fields. The most likely for CDCG purposes is the AthenaID field.

UCI also has a couple other fields that can be used as indexes (ScrambledId or PatientLocalIdentifier)

Can we make it so "MRN" isnt hardcoded into the convert.py program? If convert.py needs to know what to use as an index, maybe we can have that field be preceded by a special character in the conversion file (like a #).

— Reply to this email directly or view it on GitHubhttps://github.com/gittaylor/athenapy/issues/2.

rkelly commented 11 years ago

What are the SFMR requirements? I'm wondering if the requirements can be handled in the SFMR conversion file.

Thanks, Richard

On Wed, Feb 6, 2013, at 8:31, gittaylor wrote:

We will have to have at least 1 function that is unique to SFMR, since there are requirements for the SFMR implementation. We should create another function that mimics convertToSFMR() for other conversions. Simply copy the code to another function (e.g. convertToCDCG)...

On Tue, Feb 5, 2013 at 9:13 PM, rkelly notifications@github.com wrote:

Currently, convert.py required that the conversion file have three fields (MRN, CompletedDate, and ApptDate) (lines 117-119).

However, there are other fields that can be used as index fields. The most likely for CDCG purposes is the AthenaID field.

UCI also has a couple other fields that can be used as indexes (ScrambledId or PatientLocalIdentifier)

Can we make it so "MRN" isnt hardcoded into the convert.py program? If convert.py needs to know what to use as an index, maybe we can have that field be preceded by a special character in the conversion file (like a #).

— Reply to this email directly or view it on GitHubhttps://github.com/gittaylor/athenapy/issues/2.

Reply to this email directly or view it on GitHub: https://github.com/gittaylor/athenapy/issues/2#issuecomment-13190630

gittaylor commented 11 years ago

That would be too complicated. We would have to add extra fields. In my experience, having the ability to generate a function specific for each conversion is almost always necessary.

On Wed, Feb 6, 2013 at 5:36 PM, rkelly notifications@github.com wrote:

What are the SFMR requirements? I'm wondering if the requirements can be handled in the SFMR conversion file.

Thanks, Richard

On Wed, Feb 6, 2013, at 8:31, gittaylor wrote:

We will have to have at least 1 function that is unique to SFMR, since there are requirements for the SFMR implementation. We should create another function that mimics convertToSFMR() for other conversions. Simply copy the code to another function (e.g. convertToCDCG)...

On Tue, Feb 5, 2013 at 9:13 PM, rkelly notifications@github.com wrote:

Currently, convert.py required that the conversion file have three fields (MRN, CompletedDate, and ApptDate) (lines 117-119).

However, there are other fields that can be used as index fields. The most likely for CDCG purposes is the AthenaID field.

UCI also has a couple other fields that can be used as indexes (ScrambledId or PatientLocalIdentifier)

Can we make it so "MRN" isnt hardcoded into the convert.py program? If convert.py needs to know what to use as an index, maybe we can have that field be preceded by a special character in the conversion file (like a #).

— Reply to this email directly or view it on GitHub< https://github.com/gittaylor/athenapy/issues/2>.

Reply to this email directly or view it on GitHub: https://github.com/gittaylor/athenapy/issues/2#issuecomment-13190630

— Reply to this email directly or view it on GitHubhttps://github.com/gittaylor/athenapy/issues/2#issuecomment-13190904.