#!/usr/bin/env python3
from pathlib import Path
import re

inp = Path(r"d:\TJWaterServer\epanet\szhskeleton-patternfixed-ascii.inp")
text = inp.read_text(encoding='utf-8')
lines = text.splitlines()

start = None
for i,l in enumerate(lines):
    if l.strip().upper() == '[VALVES]':
        start = i
        break
if start is None:
    print('No [VALVES] section found')
    raise SystemExit(1)
# collect until next section header or EOF
end = len(lines)
for j in range(start+1, len(lines)):
    if re.match(r"^\s*\[.+\]", lines[j]):
        end = j
        break
block_lines = lines[start+1:end]

ids = []
for idx,l in enumerate(block_lines, start=start+1):
    if not l.strip() or l.strip().startswith(';'):
        continue
    # first token
    tok = l.split()[0]
    ids.append((idx, tok, l))

from collections import defaultdict
count = defaultdict(list)
for ln, tok, l in ids:
    count[tok].append(ln)

dups = {k:v for k,v in count.items() if len(v)>1}
print('Total valve entries found:', len(ids))
print('Duplicate token count:', len(dups))
if dups:
    print('\nSample duplicates:')
    for k,v in list(dups.items())[:20]:
        print(k, 'lines:', v)

# show whether tokens are purely digits
num_only = [tok for ln,tok,l in ids if re.fullmatch(r'\d+', tok)]
print('\nNumeric-only valve IDs count:', len(num_only))

# show examples of numeric-only
if num_only:
    print('Examples:', num_only[:20])

# write a short report
rep = inp.with_name(inp.stem + '-valves-report.txt')
with rep.open('w', encoding='utf-8') as f:
    f.write(f'Total valve entries: {len(ids)}\n')
    f.write(f'Duplicate tokens: {len(dups)}\n')
    for k,v in dups.items():
        f.write(f'{k}: lines {v}\n')
    f.write('\nNumeric-only tokens:\n')
    for tok in sorted(set(num_only)):
        f.write(tok + '\n')

print('Wrote report to', rep)