Read all metadata columns when query parsing fails

In the case that column names cannot be extracted from the query, read
all columns from the metadata. The behavior in this case is comparable
to the behavior prior to "filter: Read a subset of metadata columns"
(d0f36a1).

augur/filter/io.py

@@ -3,6 +3,7 @@
 from argparse import Namespace
 import os
 import re
+from textwrap import dedent
 from typing import Sequence, Set
 import numpy as np
 from collections import defaultdict
@@ -65,7 +66,16 @@ def get_useful_metadata_columns(args: Namespace, id_column: str, all_columns: Se
         # Attempt to automatically extract columns from the query.
         variables = extract_variables(args.query)
         if variables is None and not args.query_columns:
-            raise AugurError("Could not infer columns from the pandas query. If the query is valid, please specify columns using --query-columns.")
+            print_err(dedent(f"""\
+                WARNING: Could not infer columns from the pandas query. Reading all metadata columns,
+                which may impact execution time. If the query is valid, please open a new issue:
+
+                    <https://github.com/nextstrain/augur/issues/new/choose>
+
+                and add the query to the description:
+
+                    {args.query}"""))
+            columns.update(all_columns)
         else:
             columns.update(variables)
 

tests/functional/filter/cram/filter-query-backtick-quoting.t

@@ -12,16 +12,21 @@ Create metadata file for testing.
   > SEQ_4	
   > ~~
 
-The 'region name' column should be query-able by backtick quoting.
-This does not currently work due to a bug.
+The 'region name' column is query-able by backtick quoting.
 
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --query '(`region name` == "A")' \
   >  --output-strains filtered_strains.txt > /dev/null
-  ERROR: Could not infer columns from the pandas query. If the query is valid, please specify columns using --query-columns.
-  [2]
+  WARNING: Could not infer columns from the pandas query. Reading all metadata columns,
+  which may impact execution time. If the query is valid, please open a new issue:
+  
+      <https://github.com/nextstrain/augur/issues/new/choose>
+  
+  and add the query to the description:
+  
+      (`region name` == "A")
 
   $ sort filtered_strains.txt
-  sort: No such file or directory
-  [2]
+  SEQ_1
+  SEQ_2

tests/functional/filter/cram/filter-query-errors.t

@@ -41,5 +41,15 @@ However, other Pandas errors are not so helpful, so a link is provided for users
   >  --metadata "$TESTDIR/../data/metadata.tsv" \
   >  --query "some bad syntax" \
   >  --output-strains filtered_strains.txt > /dev/null
-  ERROR: Could not infer columns from the pandas query. If the query is valid, please specify columns using --query-columns.
+  WARNING: Could not infer columns from the pandas query. Reading all metadata columns,
+  which may impact execution time. If the query is valid, please open a new issue:
+
+      <https://github.com/nextstrain/augur/issues/new/choose>
+
+  and add the query to the description:
+
+      some bad syntax
+  ERROR: Internal Pandas error when applying query:
+  	invalid syntax (<unknown>, line 1)
+  Ensure the syntax is valid per <https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#indexing-query>.
   [2]