Facets warning appears even when facets are specified

[onnyyonn]

I am searching for all the files matching a given list of parameters. Even though I specify facets while creating a new_context, I still get the facets warning. In case it matters, I am creating a number of new contexts inside loops. Here is an example code snippet for what I am trying to do:

from pyesgf.search import SearchConnection
import numpy as np
import json, os, operator, itertools, tqdm

query_params = {"latest": True,
                "mip_era": "CMIP6",
                "activity_id": "HighResMIP",
                "realm": "atmos",
                "frequency": "6hr",
                }
variables = ['ua', 'uas', 'va', 'vas', 'psl', 'zg']
experiment_ids = ['highres-future', 'hist-1950']
source_ids = ["MPI-ESM1-2-XR", "MPI-ESM1-2-HR"]

def search_esgf(source_id, experiment_id, variable, query_params, savefile):
    try:
        ctx = conn.new_context(source_id=source_id, experiment_id=experiment_id, variable=variable, facets='source_id', **query_params)
        results = ctx.search()

        # Convert search results into a list of filename and download url
        files = []
        for i in range(0, len(results)):
            files.extend(list(map(lambda f : {'filename': f.filename, 'url': f.download_url, 'size': f.size, 'checksum': f.checksum, 'checksum_type': f.checksum_type},
                                    results[i].file_context().search())))

        # Consolidate all duplicate download links into a single entry
        files = sorted(files, key=operator.itemgetter("filename"))
        files_gr=[]
        for i,g in itertools.groupby(files, key=operator.itemgetter("filename")):
            grp = list(g)
            entry = grp[0].copy()
            entry['url'] = tuple(e['url'] for e in grp)
            files_gr.append(entry)

        # Save file list
        with open(savefile, 'w') as f:
            json.dump(files_gr, f, indent = 4)

    except KeyboardInterrupt:
        raise KeyboardInterrupt

    except Exception as e:
        print(e)

conn = SearchConnection('https://esgf-data.dkrz.de/esg-search', distrib=True)
for sid in tqdm.tqdm(source_ids, desc="Models", position=0):
    for eid in tqdm.tqdm(experiment_ids, desc="Experiments", position=1):
        for v in tqdm.tqdm(variables, desc="Variables", position=2):
            savefile = query_params["activity_id"]+"_"+eid+"_"+sid+"_"+v+"_"+query_params["frequency"]+".json"
            if not os.path.exists(savefile):
                search_esgf(sid, eid, v, query_params, savefile)

Here I am calling the search_esgf function inside loops, and inside that function a new context is being created, with facets='source_id' specified. Why do I still get the facets warning?

[siankg]

I'm having the same issue!

[JimCircadian]

@onnyyonn and @siankg the issue is the ordering, because the code is catching the additional unspecified keyword arguments internally, you need to use connection.new_context(facets="source", **query_params) rather than connection.new_context(**query, facets="source"). Make sure the parameters destined for new_context come before the search parameters.

[siankg]

Thanks for the info! I tried switching it but got the same error message.

ctx = conn.new_context(facets='project,institution_id,source_id,activity_id,experiment_id,variable,frequency', latest = True, institution_id = mod_inf[0], source_id = mod_inf[1], project = 'CMIP6', activity_id='CMIP', experiment_id = 'historical', variable='fBNF,cVeg,nVeg,npp,pr,tas', frequency='mon')

[JimCircadian]

Thanks for the info! I tried switching it but got the same error message.

ctx = conn.new_context(facets='project,institution_id,source_id,activity_id,experiment_id,variable,frequency', latest = True, institution_id = mod_inf[0], source_id = mod_inf[1], project = 'CMIP6', activity_id='CMIP', experiment_id = 'historical', variable='fBNF,cVeg,nVeg,npp,pr,tas', frequency='mon')

Gave myself a red herring (it disappeared temporarily for me, but it was late at night): will post here again, having an investigate!

[JimCircadian]

Further to this, I wasn't realising the warning comes from the invocation of file_context or aggregation_context, neither of which accept facets themselves.