deweylab / metasra-api-backend Goto Github PK

When I attempt to use the 'sampletype' flag via the API, it doesn't seem to return the subset of sampletype only--it seems to just return all of the samples.

For example:

http://metasra.biostat.wisc.edu/api/v01/samples.csv?and=CL:0000763&not=EFO:0000727,DOID:4

and

http://metasra.biostat.wisc.edu/api/v01/samples.csv?and=CL:0000763&not=EFO:0000727,DOID:4&sampletype=cell%20line

return exactly the same number of samples. Is there a way to respect the sampletype designated?

API calls aren't returning anything and the website is claiming network issues.

Hi, the website seems to be running into issues, so I wanted to re-host this myself and am running into some questions during installation:

Make sure onto_lib for python3 is in the python path, and all OBO ontology files required by onto_lib are present. (For now, the easiest way to accomplish this is to copy onto_lib and the OBO files into /build-db-script/onto_lib . You might have to edit the onto_lib python files to tell it where to look for the OBO files.)

Where are these files located, or where can they be obtained from? I don't see any reference to these files anywhere.

SRA metadata subset DB : SQLite file : a byproduct of the MetaSRA pipeline

Where is this file located?

Very excited to get this up and running--thanks for the help!

I downloaded the metasra JSON file you provide ( ), and tried to replicate the results I get via your service online. I used the following snippet to search for all samples that match a DOID term:

import json 

example_ontology = 'DOID:9119'
samples = []

with open('../datasets/metasra.v1-5.json') as json_file:
    data = json.load(json_file) 

for sample in data:
    if example_ontology in data[sample]['mapped ontology terms']:
        samples.append(sample)

print(len(samples))

For the DOID term 9119, I get 756 samples, but if I use the online service, it returns 903 samples. Do you know why I'm seeing this discrepancy, and how to address it?

It would be really useful if you could allow for logical combinations of terms. For example, if I wanted to find all samples that were either RBCs, OR Monocytes, OR Granulocytes. Currently I can make multiple queries with the API and staple them all together, but it would be great if the API/website allowed for this natively!

After posting to the API route for json.samples, I keep getting 502s. Furhtermore, when you are on the metasra website and try to enter search terms, it keeps claiming that it can't connect with the following:

Error: we had a problem processing your search. Please check your network connection, refresh the page, or try again