test_deepseek_fix.py - main - Paff/declarant

test_deepseek_fix.py
102 строки4 KB
1pash1985-gif
Unified doc list UI + AI prompt for FORM_A
05 авг 2026, 10:09
05 авг 2026, 10:09e439f08
"""Test fixes for deepseek-v4-flash reasoning token exhaustion."""
import sys, os, re, json, logging
sys.path.insert(0, '/app/web')
os.chdir('/app/web')

logging.basicConfig(level=logging.INFO, format='%(name)s - %(levelname)s - %(message)s')

from ai_parser import extract_text_from_excel, client, DEEPSEEK_MODEL, INVOICE_SYSTEM_PROMPT, _get_corrections_context

xls_path = '/app/web/uploads/reports/2026-07-27/444/sources/invoice_444_1.xls'
raw_text = extract_text_from_excel(xls_path)

# Fix 1: Strip URLs (they waste tokens and aren't useful for invoice parsing)
text_no_urls = re.sub(r'https?://\S+', '', raw_text)
# Clean up multiple spaces left behind
text_no_urls = re.sub(r'  +', ' ', text_no_urls)
text_no_urls = re.sub(r' \| $', '', text_no_urls)  # Remove trailing " | " on lines

print(f'Raw text: {len(raw_text)} chars')
print(f'URL-stripped text: {len(text_no_urls)} chars ({len(text_no_urls) - len(raw_text)} chars saved)')
url_pattern = r'https?://\S+'
print(f'URLs removed: {len(re.findall(url_pattern, raw_text))}')
print('---')

system_prompt = INVOICE_SYSTEM_PROMPT + _get_corrections_context()

# Test 1: max_tokens=16384 with original text
print('=== TEST 1: max_tokens=16384, original text ===')
try:
    response = client.chat.completions.create(
        model=DEEPSEEK_MODEL,
        messages=[
            {'role': 'system', 'content': system_prompt},
            {'role': 'user', 'content': f'Вот текст инвойса:\n\n{raw_text}'},
        ],
        temperature=0.1,
        max_tokens=16384,
        response_format={'type': 'json_object'},
    )
    fr = response.choices[0].finish_reason
    cl = len(response.choices[0].message.content or '')
    ct = response.usage.completion_tokens if response.usage else 0
    rt = getattr(response.usage.completion_tokens_details, 'reasoning_tokens', '?') if response.usage and hasattr(response.usage, 'completion_tokens_details') else '?'
    print(f'finish_reason={fr}, content_len={cl}, completion_tokens={ct}, reasoning_tokens={rt}')
    if cl > 0:
        print(f'Content preview: {(response.choices[0].message.content or "")[:200]}')
except Exception as e:
    print(f'ERROR: {e}')

print()

# Test 2: max_tokens=16384 with URL-stripped text
print('=== TEST 2: max_tokens=16384, URL-stripped text ===')
try:
    response = client.chat.completions.create(
        model=DEEPSEEK_MODEL,
        messages=[
            {'role': 'system', 'content': system_prompt},
            {'role': 'user', 'content': f'Вот текст инвойса:\n\n{text_no_urls}'},
        ],
        temperature=0.1,
        max_tokens=16384,
        response_format={'type': 'json_object'},
    )
    fr = response.choices[0].finish_reason
    cl = len(response.choices[0].message.content or '')
    ct = response.usage.completion_tokens if response.usage else 0
    rt = getattr(response.usage.completion_tokens_details, 'reasoning_tokens', '?') if response.usage and hasattr(response.usage, 'completion_tokens_details') else '?'
    print(f'finish_reason={fr}, content_len={cl}, completion_tokens={ct}, reasoning_tokens={rt}')
    if cl > 0:
        print(f'Content preview: {(response.choices[0].message.content or "")[:200]}')
except Exception as e:
    print(f'ERROR: {e}')

print()

# Test 3: max_tokens=32768 with URL-stripped text
print('=== TEST 3: max_tokens=32768, URL-stripped text ===')
try:
    response = client.chat.completions.create(
        model=DEEPSEEK_MODEL,
        messages=[
            {'role': 'system', 'content': system_prompt},
            {'role': 'user', 'content': f'Вот текст инвойса:\n\n{text_no_urls}'},
        ],
        temperature=0.1,
        max_tokens=32768,
        response_format={'type': 'json_object'},
    )
    fr = response.choices[0].finish_reason
    cl = len(response.choices[0].message.content or '')
    ct = response.usage.completion_tokens if response.usage else 0
    rt = getattr(response.usage.completion_tokens_details, 'reasoning_tokens', '?') if response.usage and hasattr(response.usage, 'completion_tokens_details') else '?'
    print(f'finish_reason={fr}, content_len={cl}, completion_tokens={ct}, reasoning_tokens={rt}')
    if cl > 0:
        content = response.choices[0].message.content or ''
        print(f'Content preview: {content[:300]}')
        parsed = json.loads(content) if content else None
        if parsed:
            print(f'Parsed OK! Supplier: {parsed.get("supplier_name", "N/A")}')
except Exception as e:
    print(f'ERROR: {e}')