🚧 add error handling

augur/merge.py

@@ -390,45 +390,59 @@ def merge_sequences(args):
     if args.quiet:
         print_info = lambda *_: None
 
-    # FIXME: restore this error handling:
-    # except subprocess.CalledProcessError as err:
-    #     raise SeqKitError(f"seqkit invocation failed") from err
-
-    with seqkit('rmdup',
-            stdin =subprocess.PIPE,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.PIPE) as process:
-        # Reversed because seqkit rmdup keeps the first entry but this command
-        # should keep the last entry.
-        # NOTE(or FIXME?): For duplicates within a file, seqkit rmdup will
-        # keep the first entry which is somewhat at odds with the "last one
-        # wins" approach used for the rest of augur merge. We could error in
-        # this case (which is what augur filter does), however that requires
-        # reading all the IDs which is unnecessary complexity for a wrapper
-        # around seqkit rmdup.
-        for s in reversed(args.sequences):
-            print_info(f"Reading sequences from {s!r}…")
-            subprocess.run([*augur, "read-file", s], stdout=process.stdin)
-
-            # Add an newline character to support FASTA files that are missing one at the end.
-            # Extraneous newline characters are stripped by seqkit.
-            print(file=process.stdin)
-
-        # Let seqkit know that there are no more sequences.
-        process.stdin.close()
-
-        print_info(f"Merging sequences and writing to {args.output_sequences!r}…")
-
-        for line in iter(process.stderr.readline, ''):
-            print_info(f"[SeqKit] {line.rstrip()}")
-
-        # Write merged sequences.
-        subprocess.run([*augur, "write-file", args.output_sequences], stdin=process.stdout)
+    try:
+        with seqkit('rmdup',
+                stdin =subprocess.PIPE,
+                stdout=subprocess.PIPE,
+                stderr=subprocess.PIPE) as process:
+            # Reversed because seqkit rmdup keeps the first entry but this command
+            # should keep the last entry.
+            # NOTE(or FIXME?): For duplicates within a file, seqkit rmdup will
+            # keep the first entry which is somewhat at odds with the "last one
+            # wins" approach used for the rest of augur merge. We could error in
+            # this case (which is what augur filter does), however that requires
+            # reading all the IDs which is unnecessary complexity for a wrapper
+            # around seqkit rmdup.
+            for s in reversed(args.sequences):
+                print_info(f"Reading sequences from {s!r}…")
+                read_process = subprocess.run([*augur, "read-file", s], stdout=process.stdin)
+                try:
+                    read_process.check_returncode()
+                except subprocess.CalledProcessError as e:
+                    # FIXME: capture augur read-file errors but not seqkit errors
+                    # print_info(f"Cannot read file {s}: {e}")
+                    pass
+
+                # Add an newline character to support FASTA files that are missing one at the end.
+                # Extraneous newline characters are stripped by seqkit.
+                print(file=process.stdin)
+
+            # Let seqkit know that there are no more sequences.
+            try:
+                process.stdin.close()
+            except BrokenPipeError:
+                # Let the return code handle this
+                pass
+
+            print_info(f"Merging sequences and writing to {args.output_sequences!r}…")
+
+            # FIXME: add error handling?
+            for line in iter(process.stderr.readline, ''):
+                print_info(f"[SeqKit] {line.rstrip()}")
+
+            # Write merged sequences.
+            # FIXME: add error handling
+            subprocess.run([*augur, "write-file", args.output_sequences], stdin=process.stdout)
 
+    finally:
         # Wait for the seqkit process to finish.
-        process.wait()
+        returncode = process.wait()
         # FIXME: address potential deadlock issue?
         # <https://docs.python.org/3/library/subprocess.html#subprocess.Popen.wait>
+        # Can't use Popen.communicate() because the data read is buffered in memory.
+        # <https://docs.python.org/3/library/subprocess.html#subprocess.Popen.communicate>
+        if returncode != 0:
+            raise SeqKitError(f"seqkit invocation failed, see error message above.")
 
 
 def sqlite3(*args, **kwargs):
@@ -517,7 +531,7 @@ def seqkit(*args, **kwargs):
     return subprocess.Popen(argv, encoding="utf-8", text=True, **kwargs)
 
 
-class SeqKitError(Exception):
+class SeqKitError(AugurError):
     pass
 
 

tests/functional/merge/cram/merge-sequences.t

@@ -60,3 +60,19 @@ This behavior is a reflection of the seqkit rmdup internals.
   ATCG
   >seq2
   GCTA
+
+Error handling
+
+  $ cat >y.fasta <<~~
+  > invalid fasta file
+  > ~~
+
+  $ ${AUGUR} merge \
+  >   --sequences x.fasta y.fasta \
+  >   --output-sequences - > merged.fasta
+  Reading sequences from 'y.fasta'…
+  Reading sequences from 'x.fasta'…
+  Merging sequences and writing to '-'…
+  [SeqKit] \x1b[31m[ERRO]\x1b[0m fastx: invalid FASTA/Q format (esc)
+  ERROR: seqkit invocation failed, see error message above.
+  [2]