fc405445c6
- Removed obsolete slots: `has_or_had_custodian_observation`, `provider`, and `specificity_annotation`. - Updated `has_or_had_score` slot to use `SpecificityScore` class and modified its description and examples. - Added new slots: `end_seconds`, `end_time`, `has_archive_path`, `has_or_had_custodian_name`, `protocol_name`, and `protocol_version`. - Introduced a script `check_annotation_types.py` to validate the presence and structure of `custodian_types` in YAML files. - Added a script `update_specificity.py` to automate updates related to `SpecificityAnnotation` to `SpecificityScore`.
89 lines
2.6 KiB
Python
89 lines
2.6 KiB
Python
import os
|
|
import networkx as nx
|
|
|
|
def get_imports(path):
|
|
imports = []
|
|
try:
|
|
with open(path, 'r') as f:
|
|
lines = f.readlines()
|
|
in_imports = False
|
|
for line in lines:
|
|
stripped = line.strip()
|
|
if stripped == "imports:":
|
|
in_imports = True
|
|
continue
|
|
if in_imports:
|
|
if not stripped.startswith("-"):
|
|
if stripped and not stripped.startswith("#"):
|
|
in_imports = False
|
|
else:
|
|
imp = stripped.lstrip("- ").strip()
|
|
imports.append(imp)
|
|
except Exception:
|
|
pass
|
|
return imports
|
|
|
|
def build_graph(root_dir):
|
|
G = nx.DiGraph()
|
|
|
|
# Map file paths to node names
|
|
# Node name: relative path from root_dir or filename if unique
|
|
# LinkML imports are relative paths.
|
|
|
|
# We need to resolve relative paths.
|
|
|
|
# files: list of (abs_path, rel_path_from_root)
|
|
# root_dir is schemas/20251121/linkml
|
|
|
|
files_map = {} # abs_path -> rel_path
|
|
|
|
for root, dirs, files in os.walk(root_dir):
|
|
for file in files:
|
|
if file.endswith(".yaml"):
|
|
abs_path = os.path.join(root, file)
|
|
rel_path = os.path.relpath(abs_path, root_dir)
|
|
files_map[abs_path] = rel_path
|
|
G.add_node(rel_path)
|
|
|
|
for abs_path, rel_path in files_map.items():
|
|
imports = get_imports(abs_path)
|
|
base_dir = os.path.dirname(abs_path)
|
|
|
|
for imp in imports:
|
|
# imp is relative to abs_path
|
|
# resolve it
|
|
if imp.startswith("linkml:"):
|
|
continue
|
|
|
|
imp_abs = os.path.normpath(os.path.join(base_dir, imp))
|
|
if not imp_abs.endswith(".yaml"):
|
|
imp_abs += ".yaml"
|
|
|
|
if imp_abs in files_map:
|
|
target = files_map[imp_abs]
|
|
G.add_edge(rel_path, target)
|
|
else:
|
|
# print(f"Warning: {rel_path} imports {imp} which resolves to {imp_abs} (not found)")
|
|
pass
|
|
|
|
return G
|
|
|
|
def find_cycles(G):
|
|
try:
|
|
cycles = nx.simple_cycles(G)
|
|
count = 0
|
|
for cycle in cycles:
|
|
print("Cycle found:")
|
|
print(cycle)
|
|
count += 1
|
|
if count >= 10:
|
|
print("Stopping after 10 cycles.")
|
|
break
|
|
if count == 0:
|
|
print("No cycles found.")
|
|
except Exception as e:
|
|
print(f"Error finding cycles: {e}")
|
|
|
|
if __name__ == "__main__":
|
|
G = build_graph("schemas/20251121/linkml")
|
|
find_cycles(G)
|