Merge pull request #2 from pineoncellar/main

feat: simple parser rules load and log processing

3 files changed, 111 insertions, 5 deletions

diff --git a/src/conventionalrp/core/parser.py b/src/conventionalrp/core/parser.py
index 32b1b9f..73f349f 100644
--- a/src/conventionalrp/core/parser.py
+++ b/src/conventionalrp/core/parser.py
@@ -1,15 +1,54 @@
+import json
+import re
+from pathlib import Path
+
 class Parser:
     def __init__(self):
         self.rules = []
 
-    def load_rules(self, rules):
+    def load_rules(self, rules_path : str):
         """Load parsing rules."""
+        if not Path(rules_path).exists():
+            raise FileNotFoundError(f"No such file or directory: {rules_path} ")
+
+        with open(rules_path, "r", encoding="utf-8") as f:
+            rules = json.load(f)
+
+        # validation rule format
+        if rules is None:
+            raise ValueError(f"Rule file cannot be empty.")
+        # to be continue...
+
         self.rules = rules
 
-    def parse_log(self, log):
+    def parse_log(self, log_path: str):
         """Parse the TRPG log based on loaded rules."""
         parsed_data = []
-        for rule in self.rules:
-            # Implement rule-based parsing logic here
-            pass
+
+        if not Path(log_path).exists():
+            raise FileNotFoundError(f"No such file or directory: {log_path} ")
+
+        with open(log_path, "r", encoding="utf-8") as f:
+            log_content = f.read().splitlines()
+
+        # Iterate each line of the log
+        for line in log_content:
+            # pass blank line
+            if not line.strip():
+                continue
+
+            # try to match the current line by rules
+            for rule in self.rules:
+                pattern = rule.get("pattern")
+                rule_type = rule.get("type")
+                match = re.search(pattern, line)
+                if match:
+                    # matched
+                    content = match.group(1).strip()
+                    parsed_data.append({"content": content, "type": rule_type})
+                    break
+            # no matched, marked as an unknown type
+            else:
+                parsed_data.append({"content": line.strip(), "type": "unknown"})
+
         return parsed_data
\ No newline at end of file
diff --git a/test/example_log.log b/test/example_log.log
new file mode 100644
index 0000000..aff1d3a
--- /dev/null
+++ b/test/example_log.log
@@ -0,0 +1,35 @@
+墨勒托.DW(1571806261) 2025-01-27 19:58:15

+#吹个口哨

+

+MIKU(2754533655) 2025-01-27 19:58:39

+她和没听到一样走远了

+

+卡梅拉(371697639) 2025-01-27 19:58:39

+什么奇怪的拟声词

+

+以实玛利(1316702392) 2025-01-27 19:58:42

+“很高兴认识你，我是以实玛利”

+

+以实玛利(1316702392) 2025-01-27 20:00:02

+（白师傅，能不能别念了）

+

+以实玛利(1316702392) 2025-01-27 20:00:07

+（变语音团了）

+

+西吉斯蒙德(1826367949) 2025-01-27 20:00:36

+（有无温州酒店推荐）

+

+以实玛利(1316702392) 2025-01-27 20:00:43

+.rc 侦查

+

+墨勒托.DW(1571806261) 2025-01-27 20:00:46

+（汉庭（

+

+MIKU(2754533655) 2025-01-27 20:01:58

+那么你在他看走廊探头探脑时

+

+MIKU(2754533655) 2025-01-27 20:02:06

+也站起身子探头探脑的看了看办公室

+

+MIKU(2754533655) 2025-01-27 20:02:18

+可以注意到瑞德曼的办公桌上有一份卷宗，卷宗上写着“4·22 袭警案调查报告”的字样：
\ No newline at end of file
diff --git a/test/example_rule.json b/test/example_rule.json
new file mode 100644
index 0000000..d385f38
--- /dev/null
+++ b/test/example_rule.json
@@ -0,0 +1,32 @@
+[

+    {

+        "pattern": "^(\\S+)\\((\\d+)\\)\\s+(\\d{4}-\\d{2}-\\d{2} \\d{2}:\\d{2}:\\d{2})",

+        "type": "metadata",

+        "description": "匹配日志元数据，提取 id、QQ 账号和时间。例如：墨勒托.DW(1571806261) 2025-01-27 19:58:15"

+    },

+    {

+        "pattern": "^#(.+)",

+        "type": "action",

+        "description": "匹配行动，以 # 开头。例如：# 我推开门"

+    },

+    {

+        "pattern": "“(.+)”",

+        "type": "speech",

+        "description": "匹配玩家发言，双引号内的内容。例如：\"你好，我是冒险者\""

+    },

+    {

+        "pattern": "\\((.+)\\)",

+        "type": "ooc_speech",

+        "description": "匹配场外发言，括号内的内容。例如：(今天没时间跑团)"

+    },

+    {

+        "pattern": "^(?:[\\.。]([^.。].+))",

+        "type": "dice_order",

+        "description": "匹配掷骰指令，以 . 或 。 开头但是不匹配连续的指令前缀。例如：匹配.ra智力，不匹配'。。。'"

+    },

+    {

+        "pattern": "【(.+)】",

+        "type": "thought",

+        "description": "匹配角色心理活动。例如：【这里好可怕】"

+    }

+]
\ No newline at end of file