识别基报数据

apps/riic_report_analysis.py

@@ -1,11 +1,14 @@
 from pepperbot.core.message.chain import MessageChain, Image, Text
-from paddleocr import PaddleOCR
+from rapidocr_onnxruntime import RapidOCR
 import os
 import asyncio
 import functools
+import re
 
 
-ocr = PaddleOCR(use_angle_cls=False, use_gpu=False)
+rapid_ocr = None
+if not rapid_ocr:
+    rapid_ocr = RapidOCR("rapid_ocr.yml")
 
 
 class RIICReportAnalysis:
@@ -14,14 +17,50 @@ class RIICReportAnalysis:
             return
         img_seg: Image = chain[0]
         img_path = await img_seg.download()
-        print(f"Image saved to {img_path}")
 
         loop = asyncio.get_running_loop()
 
-        ocr_text = await loop.run_in_executor(
-            None, functools.partial(ocr.ocr, img_path, cls=False)
+        result, elapse = await loop.run_in_executor(
+            None, functools.partial(rapid_ocr, img_path)
         )
-        print(ocr_text)
+        if not "副手简报" in [i[1] for i in result]:
+            return
+        lmb_height = 0
+        lmb_numers = []
+        date_list = []
+        exp_list = []
+        gold_list = []
+        for i in result:
+            if i[1] == "龙门币":
+                lmb_height = i[0][0][1]
+                break
+        for i in result:
+            if m := re.search(r"([0-9]+\.[0-9]+)", i[1]):
+                date_list.append([int(i[0][0][0]), m.group(1)])
+                continue
+            if m := re.search(r"EXP([0-9]+)的作战记录", i[1]):
+                exp_list.append([int(i[0][0][0]), m.group(1)])
+                continue
+            if m := re.search(r"([0-9]+)的贵金属", i[1]):
+                gold_list.append([int(i[0][0][0]), m.group(1)])
+                continue
+            if lmb_height - 20 < i[0][0][1] < lmb_height + 20:
+                if m := re.search(r"([0-9]+)", i[1]):
+                    lmb_numers.append([int(i[0][0][0]), m.group(1)])
+
+        lmb_numers.sort(key=lambda x: x[0])
+        gold_list.sort(key=lambda x: x[0])
+        exp_list.sort(key=lambda x: x[0])
+        date_list.sort(key=lambda x: x[0])
+
+        output = ""
+
+        for i in range(len(date_list)):
+            output += f"【{date_list[i][1]}】\n"
+            output += f"💵 {lmb_numers[2 * i][1]}订单({lmb_numers[2 * i + 1][1]})\n"
+            output += f"🧈 {gold_list[i][1]}赤金\n"
+            output += f"📼 {exp_list[i][1]}经验\n\n"
+
+        await chain.onebot_reply(Text(output.strip()))
 
         os.remove(img_path)
-        print(f"file removed.")