Improve branch filtering and raise error on empty dataframes

docs/use.md

@@ -41,8 +41,20 @@ github-activity jupyter/notebook -s 6.0.0 -u 6.0.1 -o sample_notebook_activity.m
 
 You can find the [resulting markdown here](sample_notebook_activity).
 
-```{tip}
-For repositories that use multiple branches, it may be necessary to filter PRs by a branch name.  This can be done using the `--branch` parameter in the CLI.   Other git references can be used as well in place of a branch name.
+## Filter pull requests by branch
+
+Many repositories work with multiple active branches (e.g., `main`, `develop`, feature branches). When generating a changelog for a specific release, you typically only want to include pull requests that were merged into the release branch.
+
+Use the `--branch` (or `-b`) parameter to filter pull requests by their target branch:
+
+```bash
+github-activity org/repo --since v1.0.0 --until v2.0.0 --branch main
+```
+
+This will **only include pull requests that targeted the `main` branch**, excluding any PRs merged to other branches like `develop` or feature branches.
+
+```{note}
+You can use any git reference (tag, commit hash, etc.) in place of a branch name.
 ```
 
 ## Choose a date or a tag to filter activity
@@ -217,7 +229,7 @@ This is not as well-documented as the CLI, but should have most functionality av
 
 For generating markdown changelogs from Python, here's an example:
 
-```
+```python
 from github_activity import generate_activity_md
 
 markdown = generate_activity_md(
@@ -231,7 +243,7 @@ markdown = generate_activity_md(
     include_opened=True,
     strip_brackets=True,
     heading_level=1,
-    branch=None,
+    branch="main",  # Filter PRs by target branch (optional, use None for all branches)
 )
 
 # Print or save the markdown

github_activity/cli.py

@@ -123,7 +123,9 @@
     "--branch",
     "-b",
     default=None,
-    help=("""The branch or reference name to filter pull requests by"""),
+    help=(
+        """Filter pull requests by their target branch. Only PRs merged into this branch will be included. """
+    ),
 )
 parser.add_argument(
     "--all",
@@ -223,31 +225,36 @@ def main():
         ignored_contributors=args.ignore_contributor,
     )
 
-    if args.all:
-        md = generate_all_activity_md(args.target, **common_kwargs)
+    # Wrap in a try/except so we don't have an ugly stack trace if there's an error
+    try:
+        if args.all:
+            md = generate_all_activity_md(args.target, **common_kwargs)
 
-    else:
-        md = generate_activity_md(
-            args.target,
-            since=args.since,
-            until=args.until,
-            heading_level=args.heading_level,
-            **common_kwargs,
-        )
+        else:
+            md = generate_activity_md(
+                args.target,
+                since=args.since,
+                until=args.until,
+                heading_level=args.heading_level,
+                **common_kwargs,
+            )
 
-    if not md:
-        return
-
-    if args.output:
-        output = os.path.abspath(args.output)
-        output_dir = os.path.dirname(output)
-        if not os.path.exists(output_dir):
-            os.makedirs(output_dir)
-        with open(args.output, "w") as ff:
-            ff.write(md)
-        print(f"Finished writing markdown to: {args.output}", file=sys.stderr)
-    else:
-        print(md)
+        if not md:
+            return
+
+        if args.output:
+            output = os.path.abspath(args.output)
+            output_dir = os.path.dirname(output)
+            if not os.path.exists(output_dir):
+                os.makedirs(output_dir)
+            with open(args.output, "w") as ff:
+                ff.write(md)
+            print(f"Finished writing markdown to: {args.output}", file=sys.stderr)
+        else:
+            print(md)
+    except ValueError as e:
+        print(f"Error: {e}", file=sys.stderr)
+        sys.exit(1)
 
 
 if __name__ == "__main__":

github_activity/github_activity.py

@@ -462,8 +462,27 @@ def generate_activity_md(
     data = get_activity(
         target, since=since, until=until, kind=kind, auth=auth, cache=False
     )
+
+    # Raise error if GitHub API returned no activity at all
+    # This happens when the repository has no issues/PRs in the date range
     if data.empty:
-        return
+        raise ValueError(
+            f"No activity found for {org}/{repo} between {since} and {until}."
+        )
+
+    # Filter the PRs by branch (or ref) if given
+    if branch is not None:
+        index_names = data[
+            (data["kind"] == "pr") & (data["baseRefName"] != branch)
+        ].index
+        data.drop(index_names, inplace=True)
+
+        # Raise error if branch filter removed all data
+        # This happens when PRs exist but none targeted the specified branch
+        if data.empty:
+            raise ValueError(
+                f"Found activity, but none for the --branch target specified for {org}/{repo} on branch '{branch}' between {since} and {until}.\n"
+            )
 
     # Collect authors of comments on issues/prs that they didn't open for our attribution list
     comment_response_cutoff = 6  # Comments on a single issue
@@ -581,15 +600,6 @@ def filter_ignored(userlist):
     ].index.tolist()
     all_contributors |= set(c for c in comment_contributors if isinstance(c, str))
 
-    # Filter the PRs by branch (or ref) if given
-    if branch is not None:
-        index_names = data[
-            (data["kind"] == "pr") & (data["baseRefName"] != branch)
-        ].index
-        data.drop(index_names, inplace=True)
-        if data.empty:
-            return
-
     # Extract datetime strings from data attributes for pandas query
     since_dt_str = data.since_dt_str  # noqa: F841
     until_dt_str = data.until_dt_str  # noqa: F841