From 70a21e3762a717377c2df2b902cf852ace4a777c Mon Sep 17 00:00:00 2001
From: Victor Lin <13424970+victorlin@users.noreply.github.com>
Date: Tue, 15 Oct 2024 11:24:08 -0700
Subject: [PATCH] Parse unnamed inputs

Preparing for sequence support, which allows unnamed inputs.
---
 augur/merge.py                               | 44 ++++++++++++++++----
 tests/functional/merge/cram/merge-metadata.t |  2 +-
 2 files changed, 37 insertions(+), 9 deletions(-)

diff --git a/augur/merge.py b/augur/merge.py
index c49e77a39..38b3eccbc 100644
--- a/augur/merge.py
+++ b/augur/merge.py
@@ -186,7 +186,7 @@ def get_metadata(
     ) -> List[NamedMetadata]:
     # Validate --metadata arguments
     try:
-        metadata = parse_named_inputs(input_metadata)
+        metadata = parse_inputs(input_metadata, require_names=True)
     except UnnamedInputError as e:
         raise AugurError(dedent(f"""\
             All metadata inputs must be assigned a name, e.g. with NAME=FILE.
@@ -558,18 +558,41 @@ def quote(s):
     return quoted if quoted else shquote('')
 
 
-def parse_named_inputs(inputs: Sequence[str]):
-    if unnamed := [x for x in inputs if "=" not in x or x.startswith("=")]:
-        raise UnnamedInputError(unnamed)
+def parse_inputs(inputs: Sequence[str], require_names: bool = False):
+    """
+    Parse inputs into tuples of (name, file).
+    name is an empty string for unnamed inputs.
+
+    If names are required, this function can raise UnnamedInputError or DuplicateInputNameError.
+    If names are optional, this function can raise InvalidNamedInputError or DuplicateInputNameError.
+    """
+    # These are only used for error checking.
+    # The original order of inputs should still be used at the end.
+    invalid_named_inputs: List[str] = []
+    named_inputs: List[str] = []
+    unnamed_inputs: List[str] = []
+    for x in inputs:
+        if x.startswith("="):
+            invalid_named_inputs.append(x)
+        elif "=" in x:
+            named_inputs.append(x)
+        else:
+            unnamed_inputs.append(x)
 
-    named_inputs = pairs(inputs)
+    if require_names:
+        if bad_inputs := [*invalid_named_inputs, *unnamed_inputs]:
+            raise UnnamedInputError(bad_inputs)
+    elif invalid_named_inputs:
+        raise InvalidNamedInputError(invalid_named_inputs)
+
+    input_pairs = pairs(inputs)
 
     if duplicate_names := [name for name, count
-                                 in count_unique(name for name, _ in named_inputs)
-                                 if count > 1]:
+                                 in count_unique(name for name, _ in input_pairs)
+                                 if name != "" and count > 1]:
         raise DuplicateInputNameError(duplicate_names)
 
-    return named_inputs
+    return input_pairs
 
 
 class UnnamedInputError(Exception):
@@ -577,6 +600,11 @@ def __init__(self, unnamed: Sequence[str]):
         self.unnamed = unnamed
 
 
+class InvalidNamedInputError(Exception):
+    def __init__(self, invalid: Sequence[str]):
+        self.invalid = invalid
+
+
 class DuplicateInputNameError(Exception):
     def __init__(self, duplicates: Sequence[str]):
         self.duplicates = duplicates
diff --git a/tests/functional/merge/cram/merge-metadata.t b/tests/functional/merge/cram/merge-metadata.t
index 87b29ad79..c2270f235 100644
--- a/tests/functional/merge/cram/merge-metadata.t
+++ b/tests/functional/merge/cram/merge-metadata.t
@@ -246,8 +246,8 @@ Metadata names are required.
   
   The following inputs were missing a name:
   
-    'x.tsv'
     '=y.tsv'
+    'x.tsv'
   
   [2]