I've worked on the batch API results a bit. It isn't as complex as feared, in that it only adds a level to the return array. It seems to be possible to use the existing parsing code (or derivation of it) and apply that to each batch query result.

Or, we just save the xlsx return:

    r = requests.post(url, data=json.dumps(query))

    try:
        r.raise_for_status()
    except requests.exceptions.HTTPError as e:
        return pd.DataFrame({"Error": [str(e)]})

    with open('testing.xlsx', 'wb') as fd:
        for chunk in r.iter_content(chunk_size=128):
            fd.write(chunk)

Another thought on this is that currently we have a fix whereby "Penta D" or "Penta_E" are forced to "PentaD", etc. And it seems like it may be better to go the other way, e.g. force "PentaD" or "Penta_E" to "Penta D" and "Penta E" depending on what the API expects. Presumably it requires %20 as the space in the URL for those markers.

Another thought on this is that currently we have a fix whereby "Penta D" or "Penta_E" are forced to "PentaD", etc. And it seems like it may be better to go the other way, e.g. force "PentaD" or "Penta_E" to "Penta D" and "Penta E" depending on what the API expects. Presumably it requires %20 as the space in the URL for those markers.

I failed to consider this. The API expects: "Penta D": "".

This consideration highlights another point I've been thinking about: if a custom DB with malformed maker names (with respect to the API) is uploaded, or a batch file with malformed marker names is used, the API query will return nothing. I am thinking of adding a check prior to the API post request and throwing a warning/error message when non-compatible marker names are found. I could see someone having a mix of compatible and non-compatible markers and still wishing to query.

Remaining to resolve:

• Update README.md with new Clastr documentation following finalization of function
• Update app 'About' for Clastr batch
• CLASTR results with multiple entries for a given line do not display #25
• Thinking wheels for tables while query is running

Other things I am missing?