Merge pull request #1256: filter: Fix support for numerical IDs with …

…both metadata and sequences

CHANGES.md

@@ -11,11 +11,13 @@
 
 * parse: Fix a bug where `--fix-dates` was always applied, with a default of `--fix-dates=monthfirst`. Now, running without `--fix-dates` will leave dates as-is. [#1247][] (@victorlin)
 * `augur.io.open_file`: Previously, the docs described a type restriction on `path_or_buffer` but it was not enforced. It has been updated to allow all I/O classes, and is enforced at run-time. [#1250][] (@victorlin)
+* filter: Fix a bug where data files consisting of only numerical strain names would not work when both `--metadata` and `--sequences` are passed. [#1256][] (@victorlin)
 
 [#1146]: https://github.com/nextstrain/augur/pull/1146
 [#1240]: https://github.com/nextstrain/augur/pull/1240
 [#1247]: https://github.com/nextstrain/augur/issues/1247
 [#1250]: https://github.com/nextstrain/augur/pull/1250
+[#1256]: https://github.com/nextstrain/augur/pull/1256
 
 ## 22.0.3 (14 June 2023)
 

augur/filter/_run.py

@@ -65,6 +65,7 @@ def run(args):
             sequence_index_path,
             sep=SEQUENCE_INDEX_DELIMITER,
             index_col=SEQUENCE_INDEX_ID_COLUMN,
+            dtype={SEQUENCE_INDEX_ID_COLUMN: "string"},
         )
 
         # Remove temporary index file, if it exists.

tests/functional/filter/cram/filter-numerical-ids.t

@@ -0,0 +1,32 @@
+Setup
+
+  $ source "$TESTDIR"/_setup.sh
+
+Create a pair of files with numerical strain IDs.
+
+  $ cat >metadata.tsv <<~~
+  > strain	col1
+  > 1	A
+  > 2	B
+  > 3	C
+  > ~~
+  $ cat >sequences.fasta <<~~
+  > >1
+  > AAAA
+  > >2
+  > AAAA
+  > >3
+  > AAAA
+  > ~~
+
+Test that nothing is filtered out due to missing sequence data.
+
+  $ ${AUGUR} filter \
+  >   --metadata metadata.tsv \
+  >   --sequences sequences.fasta \
+  >   --output-strains filtered_strains.txt \
+  >   > /dev/null 2>&1
+  $ sort filtered_strains.txt
+  1
+  2
+  3