From 2afd1934fe81ef0e2abd36d8f74468d3f320504a Mon Sep 17 00:00:00 2001
From: Charles Tapley Hoyt <cthoyt@gmail.com>
Date: Mon, 2 Dec 2024 12:22:20 +0100
Subject: [PATCH] Improve warning feedback (#52)

Store warnings in a more detailed tuple to create a better failures
summary page
---
 src/bioversions/resources/update.py | 21 +++++++++++++++------
 src/bioversions/sources/__init__.py | 23 +++++++++++++++++------
 2 files changed, 32 insertions(+), 12 deletions(-)

diff --git a/src/bioversions/resources/update.py b/src/bioversions/resources/update.py
index 9d2eca63..3957a92b 100644
--- a/src/bioversions/resources/update.py
+++ b/src/bioversions/resources/update.py
@@ -15,7 +15,7 @@
     write_export,
     write_versions,
 )
-from bioversions.sources import _iter_versions
+from bioversions.sources import FailureTuple, _iter_versions
 from bioversions.version import get_git_hash
 
 __all__ = [
@@ -48,10 +48,10 @@ def _update(force: bool):  # noqa:C901
     today = datetime.now().strftime("%Y-%m-%d")
 
     changes = False
-    errors = []
-    for bv, error in _iter_versions(use_tqdm=True):
-        if error is not None or bv is None:
-            errors.append(error)
+    failure_tuples = []
+    for bv in _iter_versions(use_tqdm=True):
+        if isinstance(bv, FailureTuple):
+            failure_tuples.append(bv)
             continue
 
         if bv.name in versions:
@@ -97,7 +97,16 @@ def _update(force: bool):  # noqa:C901
         write_export(rv)
         write_versions(rv)
 
-    FAILURES_PATH.write_text("# Errors\n\n" + "\n".join(f"- {error}" for error in errors))
+    if failure_tuples:
+        click.secho(f"Writing failure summary to {FAILURES_PATH}")
+        text = "# Summary of Errors\n\n"
+        for t in failure_tuples:
+            text += f"- {t.name} - {t.message}\n"
+        text += "\n"
+        for t in failure_tuples:
+            text += f"## {t.name}\n\nUsing class: `{t.clstype}`\n\n"
+            text += f"```python-traceback\n{t.trace}\n```\n\n"
+        FAILURES_PATH.write_text(text)
 
 
 def _log_update(bv) -> None:
diff --git a/src/bioversions/sources/__init__.py b/src/bioversions/sources/__init__.py
index 2e20aab5..2b23788c 100644
--- a/src/bioversions/sources/__init__.py
+++ b/src/bioversions/sources/__init__.py
@@ -4,8 +4,10 @@
 
 import ftplib
 import logging
+import traceback
 from collections.abc import Iterable, Mapping
 from functools import lru_cache
+from typing import NamedTuple
 
 from tqdm import tqdm
 
@@ -189,14 +191,23 @@ def get_rows(use_tqdm: bool | None = False) -> list[Bioversion]:
     """Get the rows, refreshing once per day."""
     return [
         bioversion
-        for bioversion, error in _iter_versions(use_tqdm=use_tqdm)
-        if error is None and bioversion is not None
+        for bioversion in _iter_versions(use_tqdm=use_tqdm)
+        if isinstance(bioversion, Bioversion)
     ]
 
 
+class FailureTuple(NamedTuple):
+    """Holds information about failures."""
+
+    name: str
+    clstype: str
+    message: str
+    trace: str
+
+
 def _iter_versions(
     use_tqdm: bool | None = False,
-) -> Iterable[tuple[Bioversion, None] | tuple[None, str]]:
+) -> Iterable[Bioversion | FailureTuple]:
     it = tqdm(get_getters(), disable=not use_tqdm)
 
     for cls in it:
@@ -206,10 +217,10 @@ def _iter_versions(
         except (OSError, AttributeError, ftplib.error_perm):
             msg = f"failed to resolve {cls.name}"
             tqdm.write(msg)
-            yield None, msg
+            yield FailureTuple(cls.name, cls.__name__, msg, traceback.format_exc())
         except ValueError as e:
             msg = f"issue parsing {cls.name}: {e}"
             tqdm.write(msg)
-            yield None, msg
+            yield FailureTuple(cls.name, cls.__name__, msg, traceback.format_exc())
         else:
-            yield yv, None
+            yield yv