Spaces:
Sleeping
Sleeping
File size: 3,689 Bytes
4994354 7a79e7e 4994354 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 | from __future__ import annotations
import argparse, os
from pathlib import Path
from datetime import datetime
from numpy import True_
import pandas as pd
from dotenv import dotenv_values, load_dotenv
from feature.doc_alignment import DocAlignmentModel
from evaluator2 import PairwiseRelationshipEvaluator
from io import load_json_rows, save_dataframe
from reporting import write_markdown_report
from neo4j_loader import fetch_rows_from_neo4j
from schema import load_node_schema, load_nodes_schema
load_dotenv()
print("SKIP_FIELDS =", os.getenv("SKIP_FIELDS"))
def build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(
description="Analyze pairwise sample-node property relationships"
)
parser.add_argument(
"--data",
required=False,
help="Optional path to JSON array of flattened rows. If omitted, rows are loaded from Neo4j.",
)
parser.add_argument(
"--nodes",
required=True,
help="Provide list of nodes to generate data for",
)
parser.add_argument(
"--schema",
required=False,
help="Path to YAML schema for the node",
)
parser.add_argument(
"--study",
dest="study_id",
required=False,
help="Study ID for Neo4j filtering",
)
parser.add_argument(
"--output-dir",
default="outputs",
help="Directory for analysis outputs",
)
parser.add_argument(
"--neo4j-limit",
type=int,
default=None,
help="Optional row limit when loading from Neo4j",
)
return parser
def main() -> None:
parser = build_parser()
args = parser.parse_args()
# schema = load_nodes_schema(args.schema, args.nodes)
nodes = [n.strip() for n in args.nodes.split(",") if n.strip()]
schemas = load_nodes_schema(nodes)
for schema in schemas:
if args.data:
df = load_json_rows(args.data)
else:
df = fetch_rows_from_neo4j(
schema,
study_id=args.study_id,
limit=args.neo4j_limit,
)
# doc_model = DocAlignmentModel().fit(schema)
engine = PairwiseRelationshipEvaluator(schema)
results = engine.evaluate_all_pairs(df)
output_dir = Path(os.environ["OUTPUT_DIR"])
if args.output_dir:
output_dir = Path(args.output_dir)
output_dir.mkdir(parents=True, exist_ok=True)
node_name = schema.name
study_id = args.study_id if args.study_id else "all"
timestamp = datetime.now().strftime("%Y%m%d_%H%M")
file_prefix = f"{node_name}_{study_id}_{timestamp}"
save_dataframe(results, output_dir / node_name / f"{file_prefix}_pairwise_relationships.csv")
save_dataframe(results, output_dir / node_name /f"{file_prefix}_pairwise_relationships.json")
write_markdown_report(results, output_dir / node_name / f"{file_prefix}_strong_relationships.md")
if results.empty:
print("No pairwise relationships were computed.")
return
display_cols = [
"A",
"B",
"classification",
"strength",
"predictive_strength",
"support",
"determinism",
"stability",
"doc_alignment",
"suffix_match",
"substring_match",
"heldout_accuracy",
"baseline_accuracy",
]
print(results[display_cols].head(15).to_string(index=False))
results = results.rename(columns={"heldout_accuracy": "heldout_accuracy (test)"})
print(f"\nSaved outputs to {output_dir.resolve()}")
if __name__ == "__main__":
main()
|