final

Signed-off-by: Yang Wang <[email protected]>

Author: yangw-dev

.ci/scripts/benchmark_tooling/README.md

@@ -14,13 +14,14 @@ pip install -r requirements.txt
 
 Run with default output (CLI):
 ```bash
-python3 get_benchmark_analysis_data.py --startTime "2025-06-11T00:00:00" --endTime "2025-06-17T18:00:00"
+python3 .ci/scripts/benchmark_tooling/get_benchmark_analysis_data.py --startTime "2025-06-11T00:00:00" --endTime "2025-06-17T18:00:00"
 ```
 
 Additional options:
 - `--silent`: Hide processing logs, show only results
 - `--outputType df`: Display results in DataFrame format
 - `--outputType excel --excelDir "{YOUR_LOCAL_DIRECTORY}"`: Generate Excel files (`res_private.xlsx` and `res_public.xlsx`)
+- `--outputDir "{YOUR_LOCAL_DIRECTORY}"`: Specify a custom output directory for Excel/CSV files
 
 ### Python API Usage
 
@@ -33,3 +34,18 @@ fetcher = ExecutorchBenchmarkFetcher()
 fetcher.run()
 private, public = fetcher.to_df()
 ```
+
+## analyze_benchmark_stability.py
+`analyze_benchmark_stability.py` analyzes the stability of benchmark data, comparing the results of private and public devices.
+
+### Quick Start
+Install dependencies:
+```bash
+pip install -r requirements.txt
+```
+
+```
+python .ci/scripts/benchmark_tooling/analyze_benchmark_stability.py \
+    Benchmark\ Dataset\ with\ Private\ AWS\ Devices.xlsx \
+    --reference_file Benchmark\ Dataset\ with\ Public\ AWS\ Devices.xlsx
+```

.ci/scripts/analyze_benchmark_stability.py renamed to .ci/scripts/benchmark_tooling/analyze_benchmark_stability.py

@@ -61,12 +61,11 @@
 # The abbreviations used to generate the short name for the benchmark result table
 # this is used to avoid the long table name issue when generating excel file (<=31 characters)
 ABBREVIATIONS = {
-    "samsung": "smg",
-    "galaxy": "gx",
-    "ultra": "ul",
-    "android": "and",
-    "iphone": "ip",
+    "samsung": "",
+    "galaxy": "",
+    "5g": "",
     "xnnpackq8": "xnnq8",
+    "iphone15promax": "iphone15max",
 }
 
 
@@ -539,7 +538,7 @@ def find_target_tables(self, keywords, is_private) -> List[Any]:
                 f"\n============ MATCHING Found results: {len(results)}=========\n"
             )
             for keywords, match in matchings:
-                logging.info(f"Keywords: {keywords}: matchings: {match}")
+                logging.info(f"Keywords: {keywords} {len(match)} matchings: {match}")
         return list(results.values())
 
     def normalize_string(self, s, replace="_"):

.ci/scripts/benchmark_tooling/get_benchmark_analysis_data.py

@@ -1,3 +1,6 @@
 requests>=2.32.3
 xlsxwriter>=3.2.3
 pandas>=2.3.0
+openpyxl
+tabulate
+matplotlib