From 66f7c9e973b181bd771aabdc2fd9e9db33fa7df4 Mon Sep 17 00:00:00 2001
From: Michael Boniface <m.j.boniface@soton.ac.uk>
Date: Tue, 18 Feb 2025 12:37:53 +0000
Subject: [PATCH] tidied phen init
---
phen.py | 197 +++++++++++++++++++++++++++++---------------------------
1 file changed, 101 insertions(+), 96 deletions(-)
diff --git a/phen.py b/phen.py
index fca8c08..48c7d1b 100644
--- a/phen.py
+++ b/phen.py
@@ -42,9 +42,12 @@ class PhenValidationException(Exception):
self.validation_errors = validation_errors
def construct_git_url(remote_url):
+ """Constructs a git url for github or gitlab including a PAT token environment variable"""
# check the url
parsed_url = urlparse(remote_url)
-
+
+ # if github in the URL otherwise assume it's gitlab, if we want to use others such as codeberg we'd
+ # need to update this function if the URL scheme is different.
if "github.com" in parsed_url.netloc:
# get GitHub PAT from environment variable
auth = os.getenv("ACMC_GITHUB_PAT")
@@ -78,104 +81,106 @@ def init(phen_dir, remote_url):
else:
configure=True
- # configure phen directories
- if configure:
- repo = None
- # if remote then clone the repo otherwise init a local repo
- if remote_url != None:
- # add PAT token to the URL
- git_url = construct_git_url(remote_url)
-
- # clone the repo
- repo = git.cmd.Git()
- repo.clone(git_url, phen_path)
- # open repo
- repo = git.Repo(phen_path)
-
- # check if there are any commits (new repo has no commits)
- if len(repo.branches) == 0 or repo.head.is_detached: # Handle detached HEAD (e.g., after init)
- print("The phen repository has no commits yet.")
- commit_count = 0
- else:
- # Get the total number of commits in the default branch
- commit_count = sum(1 for _ in repo.iter_commits())
- print(f"Repo has previous commits: {commit_count}")
- else:
- # local repo, create the directories and init
- phen_path.mkdir(parents=True, exist_ok=True)
- print(f"Phen directory '{phen_path}' has been created.")
- repo = git.Repo.init(phen_path)
- commit_count = 0
+ if not configure:
+ print(f"Exiting, phenotype not initiatised")
+ return
- # initialise empty repos
- if commit_count == 0:
- # create initial commit
- initial_file_path = phen_path / "README.md"
- with open(initial_file_path, "w") as file:
- file.write("# Initial commit\nThis is the first commit in the phen repository.\n")
- repo.index.add([initial_file_path])
- repo.index.commit("Initial commit")
- commit_count = 1
-
- # Checkout the phens default branch, creating it if it does not exist
- if DEFAULT_GIT_BRANCH in repo.branches:
- main_branch = repo.heads[DEFAULT_GIT_BRANCH]
- main_branch.checkout()
- else:
- main_branch = repo.create_head(DEFAULT_GIT_BRANCH)
- main_branch.checkout()
-
- # if the phen path does not contain the config file then initialise the phen type
- config_path = phen_path / CONFIG_FILE
- if not config_path.exists():
- print("Creating configuration files")
- # create codes directory
- codes_path = phen_path / CODES_DIR
- codes_path.mkdir(exist_ok=True)
- keep_path = codes_path / '.gitkeep'
- keep_path.touch(exist_ok=True)
- # create maps directory
- output_path = phen_path / OUTPUT_DIR
- output_path.mkdir(exist_ok=True)
- keep_path = output_path / '.gitkeep'
- keep_path.touch(exist_ok=True)
-
- # set initial version based on the number of commits in the repo, depending on how the repo was created
- # e.g., with a README.md, then there will be some initial commits before the phen config is added
- next_commit_count = commit_count + 1
- initial_version = f"v1.0.{next_commit_count}"
-
- # create empty phen config file
- config = {
- "concept_sets": {
- "version": initial_version,
- "omop": {
- "vocabulary_id": "",
- "vocabulary_name": "",
- "vocabulary_reference": ""
- },
- "concept_set": [
- ]
- },
- "codes": [
- ]
- }
- config_path = phen_path / CONFIG_FILE
- with open(config_path, "w", encoding="utf-8") as f:
- json.dump(config, f, indent=4)
-
- # add to repo and commit
- repo.git.add(codes_path)
- repo.git.add(output_path)
- repo.git.add(all=True)
- msg = "initialised the phen git repo."
- repo.index.commit(msg)
- print(msg)
+ # Initialise repo from local or remote
+ repo = None
+ # if remote then clone the repo otherwise init a local repo
+ if remote_url != None:
+ # add PAT token to the URL
+ git_url = construct_git_url(remote_url)
+
+ # clone the repo
+ repo = git.cmd.Git()
+ repo.clone(git_url, phen_path)
+ # open repo
+ repo = git.Repo(phen_path)
+
+ # check if there are any commits (new repo has no commits)
+ if len(repo.branches) == 0 or repo.head.is_detached: # Handle detached HEAD (e.g., after init)
+ print("The phen repository has no commits yet.")
+ commit_count = 0
else:
- print(f"Phenotype configuration files already exist")
- print(f"Phenotype initialised")
+ # Get the total number of commits in the default branch
+ commit_count = sum(1 for _ in repo.iter_commits())
+ print(f"Repo has previous commits: {commit_count}")
+ else:
+ # local repo, create the directories and init
+ phen_path.mkdir(parents=True, exist_ok=True)
+ print(f"Phen directory '{phen_path}' has been created.")
+ repo = git.Repo.init(phen_path)
+ commit_count = 0
+
+ # initialise empty repos
+ if commit_count == 0:
+ # create initial commit
+ initial_file_path = phen_path / "README.md"
+ with open(initial_file_path, "w") as file:
+ file.write("# Initial commit\nThis is the first commit in the phen repository.\n")
+ repo.index.add([initial_file_path])
+ repo.index.commit("Initial commit")
+ commit_count = 1
+
+ # Checkout the phens default branch, creating it if it does not exist
+ if DEFAULT_GIT_BRANCH in repo.branches:
+ main_branch = repo.heads[DEFAULT_GIT_BRANCH]
+ main_branch.checkout()
else:
- print(f"Phenotype not initiatised")
+ main_branch = repo.create_head(DEFAULT_GIT_BRANCH)
+ main_branch.checkout()
+
+ # if the phen path does not contain the config file then initialise the phen type
+ config_path = phen_path / CONFIG_FILE
+ if config_path.exists():
+ print(f"Phenotype configuration files already exist")
+ return
+
+ print("Creating phen directory structure and config files")
+ # create codes directory
+ codes_path = phen_path / CODES_DIR
+ codes_path.mkdir(exist_ok=True)
+ keep_path = codes_path / '.gitkeep'
+ keep_path.touch(exist_ok=True)
+
+ # create concept sets directory
+ output_path = phen_path / OUTPUT_DIR
+ output_path.mkdir(exist_ok=True)
+ keep_path = output_path / '.gitkeep'
+ keep_path.touch(exist_ok=True)
+
+ # set initial version based on the number of commits in the repo, depending on how the repo was created
+ # e.g., with a README.md, then there will be some initial commits before the phen config is added
+ next_commit_count = commit_count + 1
+ initial_version = f"v1.0.{next_commit_count}"
+
+ # create empty phen config file
+ config = {
+ "concept_sets": {
+ "version": initial_version,
+ "omop": {
+ "vocabulary_id": "",
+ "vocabulary_name": "",
+ "vocabulary_reference": ""
+ },
+ "concept_set": [
+ ]
+ },
+ "codes": [
+ ]
+ }
+ config_path = phen_path / CONFIG_FILE
+ with open(config_path, "w", encoding="utf-8") as f:
+ json.dump(config, f, indent=4)
+
+ # add to git repo and commit
+ repo.git.add(codes_path)
+ repo.git.add(output_path)
+ repo.git.add(all=True)
+ repo.index.commit("initialised the phen git repo.")
+
+ print(f"Phenotype initialised")
def validate(phen_dir):
"""Validates the phenotype directory is a git repo with standard structure"""
--
GitLab