NEW: add prefix CLI

github_activity/cli.py

@@ -1,13 +1,13 @@
 import argparse
+import copy
 import os
 import sys
-from subprocess import PIPE
-from subprocess import run
 
-from .git import _git_installed_check
+from .git import _git_get_remotes, _git_installed_check
 from .github_activity import _parse_target
 from .github_activity import generate_activity_md
 from .github_activity import generate_all_activity_md
+from .github_activity import TAGS_METADATA_BASE
 
 DESCRIPTION = "Generate a markdown changelog of GitHub activity within a date window."
 parser = argparse.ArgumentParser(description=DESCRIPTION)
@@ -117,6 +117,48 @@
     action="store_true",
     help=("""Whether to include all the GitHub tags"""),
 )
+parser.add_argument(
+    "--new-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with new features""")
+)
+parser.add_argument(
+    "--enhancement-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with enhancements""")
+)
+parser.add_argument(
+    "--bug-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with bugfixes""")
+)
+parser.add_argument(
+    "--maintenance-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with maintenance and upkeep improvements""")
+)
+parser.add_argument(
+    "--documentation-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with documentation improvements""")
+)
+parser.add_argument(
+    "--api-change-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with API and breaking changes""")
+)
+parser.add_argument(
+    "--deprecate-prefix",
+    action="extend",
+    nargs="+",
+    help=("""Prefix(es) for PRs with API and breaking changes""")
+)
 
 
 def main():
@@ -131,17 +173,12 @@ def main():
     if unknown and not args.target:
         args.target = unknown[0]
 
-    tags = args.tags.split(",") if args.tags is not None else args.tags
+    tags = args.tags.split(",") if args.tags is not None else None
     # Automatically detect the target from remotes if we haven't had one passed.
     if not args.target:
         err = "Could not automatically detect remote, and none was given."
         try:
-            out = run("git remote -v".split(), stdout=PIPE)
-            remotes = out.stdout.decode().split("\n")
-            remotes = [ii for ii in remotes if ii]
-            remotes = {
-                ii.split("\t")[0]: ii.split("\t")[1].split()[0] for ii in remotes
-            }
+            remotes = _git_get_remotes()
             if "upstream" in remotes:
                 ref = remotes["upstream"]
             elif "origin" in remotes:
@@ -159,6 +196,14 @@ def main():
         except Exception:
             raise ValueError(err)
 
+    # Build mapping from tag to tag prefixes, if given
+    prefixes = {}
+    for tag in TAGS_METADATA_BASE:
+        tag_prefixes = getattr(args, f"{tag}_prefix")
+        if tag_prefixes is None:
+            continue
+        prefixes[tag] = tag_prefixes
+
     common_kwargs = dict(
         kind=args.kind,
         auth=args.auth,
@@ -167,6 +212,8 @@ def main():
         include_opened=bool(args.include_opened),
         strip_brackets=bool(args.strip_brackets),
         branch=args.branch,
+        prefixes=prefixes
+
     )
 
     if args.all:

github_activity/git.py

@@ -1,10 +1,54 @@
+import os.path
+import re
+import shlex
 import subprocess
+from tempfile import TemporaryDirectory
+from typing import Dict, List, Tuple
 
-
-def _git_installed_check():
+def _git_installed_check() -> bool:
     cmd = ["git", "--help"]
     try:
         subprocess.check_call(cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
         return True
     except subprocess.CalledProcessError:
         return False
+
+
+def _git_get_remotes() -> Dict[str, str]:
+    out = subprocess.run("git remote -v".split(), stdout=subprocess.PIPE)
+    remotes = out.stdout.decode().split("\n")
+    remotes = [ii for ii in remotes if ii]
+    return {
+        ii.split("\t")[0]: ii.split("\t")[1].split()[0] for ii in remotes
+    }
+
+
+def _git_get_remote_sha_and_tags(repo: str, pattern: str) -> List[Tuple[str, str]]:
+    # Get the sha and tag name for each tag in the target repo
+    with TemporaryDirectory() as td:
+
+        subprocess.run(
+            shlex.split(f"git clone {repo} repo"), cwd=td
+        )
+        repo = os.path.join(td, "repo")
+        subprocess.run(shlex.split("git fetch origin --tags"), cwd=repo)
+
+        cmd = 'git log --tags --simplify-by-decoration --pretty="format:%h | %D"'
+        data = (
+            subprocess.check_output(shlex.split(cmd), cwd=repo)
+            .decode("utf-8")
+            .splitlines()
+        )
+
+    # Clean up the raw data
+    pattern = f"tag: {pattern}"
+
+    def filter(datum):
+        _, tag = datum
+        # Handle the HEAD tag if it exists
+        if "," in tag:
+            tag = tag.split(", ")[1]
+        return re.match(pattern, tag) is not None
+
+    data = [d.split(" | ") for (i, d) in enumerate(data)]
+    return [d for d in data if filter(d)]

github_activity/github_activity.py

@@ -1,15 +1,12 @@
 """Use the GraphQL api to grab issues/PRs that match a query."""
+import copy
 import datetime
 import os
 import re
 import shlex
-import subprocess
 import sys
-import urllib
-from pathlib import Path
 from subprocess import PIPE
 from subprocess import run
-from tempfile import TemporaryDirectory
 
 import dateutil
 import numpy as np
@@ -18,9 +15,9 @@
 import requests
 
 from .cache import _cache_data
+from .git import _git_get_remote_sha_and_tags
 from .graphql import GitHubGraphQlQuery
 
-
 # The tags and description to use in creating subsets of PRs
 TAGS_METADATA_BASE = {
     "new": {
@@ -178,6 +175,7 @@ def generate_all_activity_md(
     include_opened=False,
     strip_brackets=False,
     branch=None,
+    prefixes=None,
 ):
     """Generate a full markdown changelog of GitHub activity of a repo based on release tags.
 
@@ -212,40 +210,16 @@ def generate_all_activity_md(
         E.g., [MRG], [DOC], etc.
     branch : string | None
         The branch or reference name to filter pull requests by.
+    prefixes: dict | None
+        The prefixes for each tag, against which PR titles will be matched.
 
     Returns
     -------
     entry: str
         The markdown changelog entry for all of the release tags in the repo.
     """
-    # Get the sha and tag name for each tag in the target repo
-    with TemporaryDirectory() as td:
-
-        subprocess.run(
-            shlex.split(f"git clone https://github.com/{target} repo"), cwd=td
-        )
-        repo = os.path.join(td, "repo")
-        subprocess.run(shlex.split("git fetch origin --tags"), cwd=repo)
-
-        cmd = 'git log --tags --simplify-by-decoration --pretty="format:%h | %D"'
-        data = (
-            subprocess.check_output(shlex.split(cmd), cwd=repo)
-            .decode("utf-8")
-            .splitlines()
-        )
-
-    # Clean up the raw data
-    pattern = f"tag: {pattern}"
-
-    def filter(datum):
-        _, tag = datum
-        # Handle the HEAD tag if it exists
-        if "," in tag:
-            tag = tag.split(", ")[1]
-        return re.match(pattern, tag) is not None
-
-    data = [d.split(" | ") for (i, d) in enumerate(data)]
-    data = [d for d in data if filter(d)]
+    # Get the (SHA, tag) pairs for this target
+    data = _git_get_remote_sha_and_tags(f"https://github.com/{target}", pattern)
 
     # Generate a changelog entry for each version and sha range
     output = ""
@@ -276,6 +250,8 @@ def filter(datum):
             include_opened=include_opened,
             strip_brackets=strip_brackets,
             branch=branch,
+            tags=tags,
+            prefixes=prefixes,
         )
 
         if not md:
@@ -303,6 +279,7 @@ def generate_activity_md(
     strip_brackets=False,
     heading_level=1,
     branch=None,
+    prefixes=None,
 ):
     """Generate a markdown changelog of GitHub activity within a date window.
 
@@ -347,7 +324,8 @@ def generate_activity_md(
         With heading_level=2 those are increased to h2 and h3, respectively.
     branch : string | None
         The branch or reference name to filter pull requests by.
-
+    prefixes: dict | None
+        The prefixes for each tag, against which PR titles will be matched.
     Returns
     -------
     entry: str
@@ -450,13 +428,30 @@ def generate_activity_md(
     # Define categories for a few labels
     if tags is None:
         tags = TAGS_METADATA_BASE.keys()
-    if not all(tag in TAGS_METADATA_BASE for tag in tags):
+    # We expect an (improper) subset of tags
+    if not set(tags) <= TAGS_METADATA_BASE.keys():
         raise ValueError(
             "You provided an unsupported tag. Tags must be "
             f"one or more of {TAGS_METADATA_BASE.keys()}, You provided:\n"
             f"{tags}"
         )
-    tags_metadata = {key: val for key, val in TAGS_METADATA_BASE.items() if key in tags}
+    # Define per-category prefixes
+    if prefixes is None:
+        prefixes = {}
+
+    # Override default prefixes with user-specified ones
+    default_prefixes = {t: v['pre'] for t, v in TAGS_METADATA_BASE.items()}
+    prefixes = {**default_prefixes, **prefixes}
+    # Take the lowercase of the prefixes
+    prefixes = {t: [s.lower() for s in p] for t, p in prefixes.items()}
+
+    # Build tag metadata
+    tags_metadata = {}
+    for tag in tags:
+        tags_metadata[tag] = {
+            **TAGS_METADATA_BASE[tag],
+            "pre": prefixes[tag]
+        }
 
     # Initialize our tags with empty metadata
     for key, vals in tags_metadata.items():
@@ -476,7 +471,7 @@ def generate_activity_md(
         )
         # Now find PRs based on prefix
         mask_pre = closed_prs["title"].map(
-            lambda title: any(f"{ipre}:" in title for ipre in kindmeta["pre"])
+            lambda title: any(f"{ipre}:" in title.lower() for ipre in kindmeta["pre"])
         )
         mask = mask | mask_pre