[test]Evaluate model performance and accuracy with UCM (#642)

# Purpose

This PR introduces a comprehensive model validation test suite to
evaluate both performance (latency metrics) and accuracy (F1-score)
under the following three key UCM caching scenarios:

Naive: No cache hits (hit rate = 0%) — serves as the baseline with full
recomputation.
Sparse: Evaluated at hit rate = 0% to assess the performance and
accuracy gains enabled by sparsity-aware mechanisms.
Prefix Caching (PC): Evaluated across multiple hit rates [0%, 30%, 50%,
80%, 100%] to demonstrate the impact of prefix reuse on inference
latency.

# Modifications 

Added test cases to verify whether the model is compatible with PC and
sparsification.


# Test

```
==============================================================================================================
Hit Rate (%) Input Tokens    Output Tokens   Concurrency  TTFT_mean [s]        TPOT_mean [s]        E2E_mean [s]        
--------------------------------------------------------------------------------------------------------------
0            8000            200             8            2.9202               0.0300               8.9285              
30           8000            200             8            2.8951               0.0297               8.8281              
50           8000            200             8            2.8731               0.0300               8.8691              
80           8000            200             8            2.9001               0.0299               8.8861              
100          8000            200             8            2.8534               0.0305               8.9540              
==============================================================================================================

========================================
Test            f1-score
----------------------------------------
PC              0.0229      
========================================
```
<!--
CI passed with new added/existing test.
If it was tested in a way different from regular unit tests, please
clarify how you tested step by step, ideally copy and paste-able, so
that other reviewers can test and check, and descendants can verify in
the future.
If tests were not added, please describe why they were not added and/or
why it was difficult to add.
-->

Co-authored-by: yuanzhg078 <939526371@qq.com>
Co-authored-by: Mag1c.H <hemajun815@163.com>

Author: 3 people

.pre-commit-config.yaml

@@ -3,6 +3,7 @@ repos:
     rev: v2.4.1
     hooks:
       - id: codespell
+        exclude: \.(jsonl|txt)$
         args: [
           '--skip', 'ucm/csrc/**,./ucm.egg-info/**,.github/**',
           '-L', 'CANN,cann,NNAL,nnal,ASCEND,ascend,EnQue,CopyIn'
@@ -22,6 +23,7 @@ repos:
     rev: v1.7.7
     hooks:
       - id: actionlint
+        exclude: \.(jsonl|txt)$
 default_stages:
   - pre-commit
   - manual

test/common/db_utils.py

@@ -3,7 +3,10 @@
 import threading
 from datetime import datetime
 from pathlib import Path
-from typing import Any, Dict, Optional
+from typing import Any, Dict, List, Optional
+
+from peewee import AutoField, DateTimeField, Model, TextField
+from playhouse.reflection import Introspector
 
 # Lazy imports for database components
 peewee = None
@@ -204,6 +207,55 @@ def write_to_db(table_name: str, data: Dict[str, Any]) -> bool:
         return False
 
 
+def read_from_db(
+    table_name: str, filters: Optional[Dict[str, Any]] = None, limit: int = 1
+) -> List[Dict[str, Any]]:
+    db_config = _get_db_config()
+    if not db_config.get("enabled", False):
+        logger.warning("Database disabled. Skipping read.")
+        return []
+
+    db = _get_db()
+    if db is None:
+        logger.error("Failed to connect to database.")
+        return []
+
+    _ensure_peewee_imported()
+
+    try:
+        introspector = Introspector.from_database(db)
+        DynamicModel = introspector.generate_models(table_names=[table_name]).get(
+            table_name
+        )
+
+        if DynamicModel is None:
+            logger.warning(f"Table '{table_name}' not found in database.")
+            return []
+
+        query = DynamicModel.select()
+
+        if filters:
+            for key, value in filters.items():
+                if hasattr(DynamicModel, key):
+                    field = getattr(DynamicModel, key)
+                    query = query.where(field == value)
+                else:
+                    logger.warning(
+                        f"Filter key '{key}' does not exist in table '{table_name}'. Skipped."
+                    )
+
+        query = query.order_by(DynamicModel.created_at.desc()).limit(limit)
+
+        results = []
+        for row in query:
+            results.append(row.__data__)
+        return results
+
+    except Exception as e:
+        logger.error(f"Error reading from table '{table_name}': {e}")
+        return []
+
+
 def database_connection(build_id: str) -> None:
     logger.info(f"Setting test build ID: {build_id}")
     _set_test_build_id(build_id)