feat: add MongoDB test generation and update dependencies

- Added pymongo==3.13.0 to requirements.txt for MongoDB connectivity
- Implemented generate_summarization_from_mongo.py script to generate summarization tests from MongoDB
- Updated run.sh to support 'gen-mongo' command for MongoDB test generation
- Enhanced scripts/README.md with documentation for new MongoDB functionality
- Improved help text in run.sh to clarify available commands and usage examples
```

This commit adds MongoDB integration for test generation and updates the documentation and scripts accordingly.

src/benchmarks/base.py

@@ -6,6 +6,9 @@ from typing import Dict, Any, List
 from abc import ABC, abstractmethod
 from models.ollama_client import OllamaClient
 
+# Импортируем константы
+from constants import TEST_SEPARATOR
+
 class Benchmark(ABC):
     """Базовый класс для всех бенчмарков."""
 

src/benchmarks/codegen.py

@@ -2,9 +2,9 @@ import logging
 import json
 import os
 from typing import Dict, Any, List
-from benchmarks.base import Benchmark
+from benchmarks.base import Benchmark, TEST_SEPARATOR
 
-class CodeGenBenchmark(Benchmark):
+class CodegenBenchmark(Benchmark):
     """Бенчмарк для тестирования генерации кода."""
 
     def __init__(self):
@@ -21,14 +21,17 @@ class CodeGenBenchmark(Benchmark):
         data_dir = "tests/codegen"
 
         for filename in os.listdir(data_dir):
-            if filename.endswith('.json'):
+            if filename.endswith('.txt'):
                 with open(os.path.join(data_dir, filename), 'r', encoding='utf-8') as f:
-                    data = json.load(f)
-                    test_data.append({
-                        'name': filename.replace('.json', ''),
-                        'prompt': data['prompt'],
-                        'expected': data['expected']
-                    })
+                    content = f.read()
+                    # Разделяем по разделителю
+                    parts = content.split(TEST_SEPARATOR, 1)
+                    if len(parts) == 2:
+                        test_data.append({
+                            'name': filename.replace('.txt', ''),
+                            'prompt': parts[0],
+                            'expected': parts[1]
+                        })
 
         return test_data
 

src/benchmarks/summarization.py

@@ -2,7 +2,7 @@ import logging
 import json
 import os
 from typing import Dict, Any, List
-from benchmarks.base import Benchmark
+from benchmarks.base import Benchmark, TEST_SEPARATOR
 
 class SummarizationBenchmark(Benchmark):
     """Бенчмарк для тестирования пересказов."""
@@ -21,14 +21,17 @@ class SummarizationBenchmark(Benchmark):
         data_dir = "tests/summarization"
 
         for filename in os.listdir(data_dir):
-            if filename.endswith('.json'):
+            if filename.endswith('.txt'):
                 with open(os.path.join(data_dir, filename), 'r', encoding='utf-8') as f:
-                    data = json.load(f)
-                    test_data.append({
-                        'name': filename.replace('.json', ''),
-                        'prompt': data['prompt'],
-                        'expected': data['expected']
-                    })
+                    content = f.read()
+                    # Разделяем по разделителю
+                    parts = content.split(TEST_SEPARATOR, 1)
+                    if len(parts) == 2:
+                        test_data.append({
+                            'name': filename.replace('.txt', ''),
+                            'prompt': parts[0],
+                            'expected': parts[1]
+                        })
 
         return test_data
 

src/benchmarks/translation.py

@@ -2,7 +2,7 @@ import logging
 import json
 import os
 from typing import Dict, Any, List
-from benchmarks.base import Benchmark
+from benchmarks.base import Benchmark, TEST_SEPARATOR
 
 class TranslationBenchmark(Benchmark):
     """Бенчмарк для тестирования переводов."""
@@ -21,14 +21,17 @@ class TranslationBenchmark(Benchmark):
         data_dir = "tests/translation"
 
         for filename in os.listdir(data_dir):
-            if filename.endswith('.json'):
+            if filename.endswith('.txt'):
                 with open(os.path.join(data_dir, filename), 'r', encoding='utf-8') as f:
-                    data = json.load(f)
-                    test_data.append({
-                        'name': filename.replace('.json', ''),
-                        'prompt': data['prompt'],
-                        'expected': data['expected']
-                    })
+                    content = f.read()
+                    # Разделяем по разделителю
+                    parts = content.split(TEST_SEPARATOR, 1)
+                    if len(parts) == 2:
+                        test_data.append({
+                            'name': filename.replace('.txt', ''),
+                            'prompt': parts[0],
+                            'expected': parts[1]
+                        })
 
         return test_data
 

src/constants.py

@@ -0,0 +1,4 @@
+"""Константы для проекта."""
+
+# Константа для разделителя в TXT файлах
+TEST_SEPARATOR = "\n==============\n"

src/main.py

@@ -4,7 +4,7 @@ from typing import List
 from models.ollama_client import OllamaClient
 from benchmarks.translation import TranslationBenchmark
 from benchmarks.summarization import SummarizationBenchmark
-from benchmarks.codegen import CodeGenBenchmark
+from benchmarks.codegen import CodegenBenchmark
 from utils.report import ReportGenerator
 
 def setup_logging(verbose: bool = False):
@@ -33,7 +33,7 @@ def run_benchmarks(ollama_client: OllamaClient, model_name: str, benchmarks: Lis
     benchmark_classes = {
         'translation': TranslationBenchmark,
         'summarization': SummarizationBenchmark,
-        'codegen': CodeGenBenchmark
+        'codegen': CodegenBenchmark
     }
 
     results = []