Merge branch 'magicpdf:master' into master

.github/workflows/benchmark.yml

@@ -18,14 +18,16 @@ jobs:
       fail-fast: true
 
     steps:
+    - name: config-net
+      run: |
+        export http_proxy=http://bigdata_open_proxy:H89k5qwQRDYfz@10.140.90.20:10811
+        export https_proxy=http://bigdata_open_proxy:H89k5qwQRDYfz@10.140.90.20:10811
     - name: PDF benchmark
       uses: actions/checkout@v3
       with:
         fetch-depth: 2
     - name: check-requirements
       run: |
-        export http_proxy=http://bigdata_open_proxy:H89k5qwQRDYfz@10.140.90.20:10811
-        export https_proxy=http://bigdata_open_proxy:H89k5qwQRDYfz@10.140.90.20:10811
         changed_files=$(git diff --name-only -r HEAD~1 HEAD)
         echo $changed_files
         if [[ $changed_files =~ "requirements.txt" ]]; then
@@ -36,7 +38,7 @@ jobs:
     - name: benchmark
       run: |
         echo "start test"
-        cd tools && python ocr_badcase.py pdf_json_label_0306.json ocr_dataset.json json_files.zip output.json
+        cd tools && python ocr_badcase.py pdf_json_label_0306.json ocr_dataset.json json_files.zip badcase.json overall.json base_data.json
   notify_to_feishu:
     if: ${{ always() && !cancelled() && contains(needs.*.result, 'failure') && (github.ref_name == 'master') }}
     needs: [pdf-test]

tools/base_data.json

+{
+    "accuracy": 1.0,
+    "precision": 1.0,
+    "recall": 1.0,
+    "f1_score": 1.0,
+    "pdf间的平均编辑距离": 133.10256410256412,
+    "pdf间的平均bleu": 0.28838311595434046,
+    "分段准确率": 0.07220216606498195,
+    "行内公式准确率": {
+        "accuracy": 0.004835727492533068,
+        "precision": 0.008790072388831437,
+        "recall": 0.010634970284641852,
+        "f1_score": 0.009624911535739562
+    },
+    "行内公式编辑距离": 1.6176470588235294,
+    "行内公式bleu": 0.17154724654721457,
+    "行间公式准确率": {
+        "accuracy": 0.08490566037735849,
+        "precision": 0.1836734693877551,
+        "recall": 0.13636363636363635,
+        "f1_score": 0.1565217391304348
+    },
+    "行间公式编辑距离": 113.22222222222223,
+    "行间公式bleu": 0.2531053359913409,
+    "丢弃文本准确率": {
+        "accuracy": 0.00035398230088495576,
+        "precision": 0.0006389776357827476,
+        "recall": 0.0007930214115781126,
+        "f1_score": 0.0007077140835102619
+    },
+    "丢弃文本标签准确率": {
+        "color_background_header_txt_block": {
+            "precision": 0.0,
+            "recall": 0.0,
+            "f1-score": 0.0,
+            "support": 41.0
+        },
+        "header": {
+            "precision": 0.0,
+            "recall": 0.0,
+            "f1-score": 0.0,
+            "support": 4.0
+        },
+        "footnote": {
+            "precision": 1.0,
+            "recall": 0.009708737864077669,
+            "f1-score": 0.019230769230769232,
+            "support": 103.0
+        },
+        "on-table": {
+            "precision": 0.0,
+            "recall": 0.0,
+            "f1-score": 0.0,
+            "support": 665.0
+        },
+        "rotate": {
+            "precision": 0.0,
+            "recall": 0.0,
+            "f1-score": 0.0,
+            "support": 63.0
+        },
+        "on-image": {
+            "precision": 0.0,
+            "recall": 0.0,
+            "f1-score": 0.0,
+            "support": 380.0
+        },
+        "micro avg": {
+            "precision": 1.0,
+            "recall": 0.0007961783439490446,
+            "f1-score": 0.0015910898965791568,
+            "support": 1256.0
+        }
+    },
+    "丢弃图片准确率": {
+        "accuracy": 0.0,
+        "precision": 0.0,
+        "recall": 0.0,
+        "f1_score": 0.0
+    },
+    "丢弃表格准确率": {
+        "accuracy": 0.0,
+        "precision": 0.0,
+        "recall": 0.0,
+        "f1_score": 0.0
+    }
+}
\ No newline at end of file