feat: Allow creation of PRs upon wrapper updates (#111)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Pull request creation added for environment and wrapper updates with
CLI control.
  * Per-file (per-Snakefile) PR option for finer-grained changes.
  * Optional automatic labeling of PRs for updated items.
* Unified and more reliable PR automation powering the above behaviors.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Author: johanneskoester

snakedeploy/client.py

@@ -166,6 +166,30 @@ def get_parser():
         help="Only warn if conda env evaluation fails and go on with the other envs.",
     )
 
+    def add_create_pr_args(subparser, entity: str):
+        subparser.add_argument(
+            "--create-prs",
+            action="store_true",
+            help=f"Create pull request for each updated {entity}. "
+            "Requires GITHUB_TOKEN and GITHUB_REPOSITORY (the repo name) and one of GITHUB_REF_NAME or GITHUB_BASE_REF "
+            "(preferring the latter, representing the target branch, e.g. main or master) environment "
+            "variables to be set (the latter three are available when running as github action). "
+            "In order to enable further actions (e.g. checks) to run on the generated PRs, the "
+            "provided GITHUB_TOKEN may not be the default github actions token. See here for "
+            "options: https://github.com/peter-evans/create-pull-request/blob/main/docs/concepts-guidelines.md#triggering-further-workflow-runs.",
+        )
+        subparser.add_argument(
+            "--entity-regex",
+            help=f"Regular expression for deriving an entity name from the {entity} file name "
+            "(will be used for adding a label and for title and description). Has to contain a group 'entity' "
+            "(e.g. '(?P<entity>.+)/environment.yaml').",
+        )
+        subparser.add_argument(
+            "--pr-add-label",
+            action="store_true",
+            help="Add a label to the PR. Has to be used in combination with --entity-regex.",
+        )
+
     update_conda_envs = subparsers.add_parser(
         "update-conda-envs",
         help="Update given conda environment definition files (in YAML format) "
@@ -187,28 +211,7 @@ def get_parser():
         action="store_true",
         help="also pin the updated environments (see pin-conda-envs subcommand).",
     )
-    update_conda_envs.add_argument(
-        "--create-prs",
-        action="store_true",
-        help="Create pull request for each updated environment. "
-        "Requires GITHUB_TOKEN and GITHUB_REPOSITORY (the repo name) and one of GITHUB_REF_NAME or GITHUB_BASE_REF "
-        "(preferring the latter, representing the target branch, e.g. main or master) environment "
-        "variables to be set (the latter three are available when running as github action). "
-        "In order to enable further actions (e.g. checks) to run on the generated PRs, the "
-        "provided GITHUB_TOKEN may not be the default github actions token. See here for "
-        "options: https://github.com/peter-evans/create-pull-request/blob/main/docs/concepts-guidelines.md#triggering-further-workflow-runs.",
-    )
-    update_conda_envs.add_argument(
-        "--entity-regex",
-        help="Regular expression for deriving an entity name from the environment file name "
-        "(will be used for adding a label and for title and description). Has to contain a group 'entity' "
-        "(e.g. '(?P<entity>.+)/environment.yaml').",
-    )
-    update_conda_envs.add_argument(
-        "--pr-add-label",
-        action="store_true",
-        help="Add a label to the PR. Has to be used in combination with --entity-regex.",
-    )
+    add_create_pr_args(update_conda_envs, "environment")
     update_conda_envs.add_argument(
         "--warn-on-error",
         action="store_true",
@@ -223,6 +226,12 @@ def get_parser():
     update_snakemake_wrappers.add_argument(
         "snakefiles", nargs="+", help="Paths to Snakefiles which should be updated."
     )
+    add_create_pr_args(update_snakemake_wrappers, "snakefile")
+    update_snakemake_wrappers.add_argument(
+        "--per-snakefile-prs",
+        action="store_true",
+        help="Create one PR per Snakefile instead of a single PR for all.",
+    )
 
     scaffold_snakemake_plugin = subparsers.add_parser(
         "scaffold-snakemake-plugin",
@@ -312,7 +321,13 @@ def help(return_code=0):
                 warn_on_error=args.warn_on_error,
             )
         elif args.subcommand == "update-snakemake-wrappers":
-            update_snakemake_wrappers(args.snakefiles)
+            update_snakemake_wrappers(
+                args.snakefiles,
+                create_prs=args.create_prs,
+                per_snakefile_prs=args.per_snakefile_prs,
+                entity_regex=args.entity_regex,
+                pr_add_label=args.pr_add_label,
+            )
         elif args.subcommand == "scaffold-snakemake-plugin":
             scaffold_plugin(args.plugin_type)
     except UserError as e:

snakedeploy/conda.py

@@ -1,24 +1,20 @@
-from collections import namedtuple
 import copy
 import json
-import os
 from pathlib import Path
 import subprocess as sp
 import tempfile
 import re
 from glob import glob
 from itertools import chain
-import github
-from urllib3.util.retry import Retry
 import random
+from typing import Optional
 
 from packaging import version as packaging_version
 import yaml
-from github import Github, GithubException
-from reretry import retry
 
 from snakedeploy.exceptions import UserError
 from snakedeploy.logger import logger
+from snakedeploy.prs import PR, get_repo
 from snakedeploy.utils import YamlDumper
 from snakedeploy.conda_version import VersionOrder
 
@@ -66,9 +62,6 @@ def update_conda_envs(
     )
 
 
-File = namedtuple("File", "path, content, is_updated, msg")
-
-
 class CondaEnvProcessor:
     def __init__(self, conda_frontend="mamba"):
         self.conda_frontend = conda_frontend
@@ -84,22 +77,16 @@ def __init__(self, conda_frontend="mamba"):
     def process(
         self,
         conda_env_paths,
-        create_prs=False,
-        update_envs=True,
-        pin_envs=True,
-        pr_add_label=False,
-        entity_regex=None,
-        warn_on_error=False,
+        create_prs: bool = False,
+        update_envs: bool = True,
+        pin_envs: bool = True,
+        pr_add_label: bool = False,
+        entity_regex: Optional[str] = None,
+        warn_on_error: bool = False,
     ):
         repo = None
         if create_prs:
-            g = Github(
-                os.environ["GITHUB_TOKEN"],
-                retry=Retry(
-                    total=10, status_forcelist=(500, 502, 504), backoff_factor=0.3
-                ),
-            )
-            repo = g.get_repo(os.environ["GITHUB_REPOSITORY"]) if create_prs else None
+            repo = get_repo()
         conda_envs = list(chain.from_iterable(map(glob, conda_env_paths)))
         random.shuffle(conda_envs)
 
@@ -109,19 +96,6 @@ def process(
             )
         for conda_env_path in conda_envs:
             if create_prs:
-                entity = conda_env_path
-                if entity_regex is not None:
-                    m = re.match(entity_regex, conda_env_path)
-                    if m is None:
-                        raise UserError(
-                            f"Given --entity-regex did not match any {conda_env_path}."
-                        )
-                    try:
-                        entity = m.group("entity")
-                    except IndexError:
-                        raise UserError(
-                            "No group 'entity' found in given --entity-regex."
-                        )
                 if pr_add_label and not entity_regex:
                     raise UserError(
                         "Cannot add label to PR without --entity-regex specified."
@@ -132,11 +106,12 @@ def process(
                 )
                 mode = "bump" if update_envs else "pin"
                 pr = PR(
-                    f"perf: auto{mode} {entity}",
-                    f"Automatic {mode} of {entity}.",
-                    f"auto{mode}/{entity.replace('/', '-')}",
+                    f"perf: auto{mode} {conda_env_path}",
+                    f"Automatic {mode} of {conda_env_path}.",
+                    f"auto{mode}/{conda_env_path.replace('/', '-')}",
                     repo,
-                    label=entity if pr_add_label else None,
+                    entity=conda_env_path,
+                    label_entity_regex=entity_regex if pr_add_label else None,
                 )
             else:
                 pr = None
@@ -161,6 +136,7 @@ def process(
                 else:
                     raise UserError(msg)
             if create_prs:
+                assert pr is not None
                 pr.create()
 
     def update_env(
@@ -303,83 +279,3 @@ def exec_conda(self, subcmd):
             universal_newlines=True,
             check=True,
         )
-
-
-class PR:
-    def __init__(self, title, body, branch, repo, label=None):
-        self.title = title
-        self.body = body
-        self.files = []
-        self.branch = branch
-        self.repo = repo
-        self.base_ref = (
-            os.environ.get("GITHUB_BASE_REF") or os.environ["GITHUB_REF_NAME"]
-        )
-        self.label = label
-
-    def add_file(self, filepath, content, is_updated, msg):
-        self.files.append(File(str(filepath), content, is_updated, msg))
-
-    @retry(tries=2, delay=60)
-    def create(self):
-        if not self.files:
-            logger.info("No files to commit.")
-            return
-
-        branch_exists = False
-        try:
-            b = self.repo.get_branch(self.branch)
-            logger.info(f"Branch {b} already exists.")
-            branch_exists = True
-        except GithubException as e:
-            if e.status != 404:
-                raise e
-            logger.info(f"Creating branch {self.branch}...")
-            self.repo.create_git_ref(
-                ref=f"refs/heads/{self.branch}",
-                sha=self.repo.get_branch(self.base_ref).commit.sha,
-            )
-        for file in self.files:
-            sha = None
-            if branch_exists:
-                logger.info(f"Obtaining sha of {file.path} on branch {self.branch}...")
-                try:
-                    # try to get sha if file exists
-                    sha = self.repo.get_contents(file.path, self.branch).sha
-                except github.GithubException.UnknownObjectException as e:
-                    if e.status != 404:
-                        raise e
-            elif file.is_updated:
-                logger.info(
-                    f"Obtaining sha of {file.path} on branch {self.base_ref}..."
-                )
-                sha = self.repo.get_contents(file.path, self.base_ref).sha
-
-            if sha is not None:
-                self.repo.update_file(
-                    file.path,
-                    file.msg,
-                    file.content,
-                    sha,
-                    branch=self.branch,
-                )
-            else:
-                self.repo.create_file(
-                    file.path, file.msg, file.content, branch=self.branch
-                )
-
-        pr_exists = any(
-            pr.head.label.split(":", 1)[1] == self.branch
-            for pr in self.repo.get_pulls(state="open", base=self.base_ref)
-        )
-        if pr_exists:
-            logger.info("PR already exists.")
-        else:
-            pr = self.repo.create_pull(
-                title=self.title,
-                body=self.body,
-                head=self.branch,
-                base=self.base_ref,
-            )
-            pr.add_to_labels(self.label)
-            logger.info(f"Created PR: {pr.html_url}")