Do type conversions on a copy of the metadata

This fixes the unexpected behavior described in the previous commit.

augur/filter/include_exclude_rules.py

@@ -184,11 +184,14 @@ def filter_by_query(metadata, query) -> FilterFunctionReturn:
     set()
 
     """
+    # Create a copy to prevent modification of the original DataFrame.
+    metadata_copy = metadata.copy()
+
     # Try converting all columns to numeric.
-    for column in metadata.columns:
-        metadata[column] = pd.to_numeric(metadata[column], errors='ignore')
+    for column in metadata_copy.columns:
+        metadata_copy[column] = pd.to_numeric(metadata_copy[column], errors='ignore')
 
-    return set(metadata.query(query).index.values)
+    return set(metadata_copy.query(query).index.values)
 
 
 def filter_by_ambiguous_date(metadata, date_column, ambiguity) -> FilterFunctionReturn:

tests/functional/filter/cram/filter-query-and-exclude-ambiguous-dates-by.t

@@ -13,12 +13,14 @@ Create metadata TSV file for testing.
   > ~~
 
 Confirm that `--exclude-ambiguous-dates-by` works for all year only ambiguous dates.
-This currently fails because the metadata DataFrame is modified in-place.
 
   $ ${AUGUR} filter \
   >  --metadata metadata.tsv \
   >  --query 'region=="Asia"' \
   >  --exclude-ambiguous-dates-by any \
   >  --empty-output-reporting silent \
-  >  --output-strains filtered_strains.txt > /dev/null 2>&1
-  [2]
+  >  --output-strains filtered_strains.txt
+  4 strains were dropped during filtering
+  \t1 of these were filtered out by the query: "region=="Asia"" (esc)
+  \t3 of these were dropped because of their ambiguous date in any (esc)
+  0 strains passed all filters