工程名称下拉项获取兼容.md文件，同时新增自定义答案合成类

backend/app/engine/loaders/__init__.py

@@ -3,7 +3,7 @@ import yaml
 from app.engine.loaders.db import DBLoaderConfig, get_db_documents
 from app.engine.loaders.file import FileLoaderConfig, get_file_documents
 from app.engine.loaders.web import WebLoaderConfig, get_web_documents
-from app.engine.loaders.projectJson import getProjectName
+from app.engine.loaders.file import getProjectName
 import os
 
 

backend/app/engine/loaders/file.py

@@ -6,6 +6,9 @@ from llama_index.core.readers.base import BaseReader
 from llama_index.core.readers.json import JSONReader
 from llama_parse import LlamaParse
 from pydantic import BaseModel, validator
+from app.engine.loaders.markdownReader import ChunkMarkdownReader
+from app.engine.loaders.projectJson import ProjectJson
+
 
 logger = logging.getLogger(__name__)
 
@@ -20,7 +23,6 @@ class FileLoaderConfig(BaseModel):
             raise ValueError(f"Directory '{v}' does not exist")
         return v
 
-
 def llama_parse_parser():
     if os.getenv("LLAMA_CLOUD_API_KEY") is None:
         raise ValueError(
@@ -35,7 +37,6 @@ def llama_parse_parser():
     )
     return parser
 
-
 def llama_parse_extractor() -> Dict[str, LlamaParse]:
     from llama_parse.utils import SUPPORTED_FILE_TYPES
 
@@ -43,8 +44,11 @@ def llama_parse_extractor() -> Dict[str, LlamaParse]:
     return {file_type: parser for file_type in SUPPORTED_FILE_TYPES}
 
 def llama_local_extractor() -> Dict[str, BaseReader]:
-    return {".json" : JSONReader(clean_json=False,levels_back=0)}
-
+    parser = {
+        ".json" : JSONReader(clean_json=False,levels_back=0),
+        ".md" : ChunkMarkdownReader(),
+        }
+    return parser
 
 def get_file_documents(config: FileLoaderConfig,childPath: str):
     from llama_index.core.readers import SimpleDirectoryReader
@@ -86,3 +90,32 @@ def get_file_documents(config: FileLoaderConfig,childPath: str):
         else:
             # Raise the error if it is not the case of empty data dir
             raise e
+
+def prjFileSuffix(dir:str):
+    entries = os.listdir(dir)
+    file_names = [entry for entry in entries if os.path.isfile(os.path.join(dir, entry))]
+    if len(file_names) > 0:
+        return os.path.splitext(file_names[0])[1]
+    return ''
+
+def getProjectName(dir:str):
+    suffix = prjFileSuffix(dir)
+    if suffix== '.json':
+        prjJson = ProjectJson(dir)
+        prjJson.parse()
+        tb = prjJson.table('工程属性')
+        records = tb.records()
+        for record in records:
+            name = record.value('名称') 
+            if name == '工程名称':
+                return record.value('值') 
+    elif suffix == '.md':
+        md_files = [f for f in os.listdir(dir) if f.endswith('.md')]
+        for md_file in md_files:
+            prjPath = os.path.join(dir, md_file)
+            basename = os.path.splitext(md_file)[0]
+            if basename =='工程属性':
+                rd = ChunkMarkdownReader()
+                rd.load_data(prjPath)
+                return rd.findValue("名称=='工程名称'",'值')
+    return ''

backend/app/engine/loaders/markdownReader.py

@@ -13,6 +13,8 @@ class ChunkMarkdownReader(MarkdownReader):
     ) -> None:
         self._chunkSize = chunkSize
         self._tokenizer = get_tokenizer()
+        self._colheader = ''
+        self._rows = []
         super().__init__(*args,**kwargs)
 
     def markdown_to_tups(self, markdown_text: str) -> List[Tuple[Optional[str], str]]:
@@ -34,6 +36,8 @@ class ChunkMarkdownReader(MarkdownReader):
                 tokensNum =  headerSize
                 current_lines.clear()
             current_lines.append(line)
+            if strTitle!='' and strheader!='':
+                self._rows.append(line)
 
             if line == '\n' or line == '\r':
                 if tokensNum > self._chunkSize:
@@ -43,10 +47,12 @@ class ChunkMarkdownReader(MarkdownReader):
                 current_lines.clear()
                         
             if line.startswith("|---"):
+                self._colheader = current_lines[0]
                 strheader = "\n".join(current_lines) 
                 headerSize= headerSize + self._token_size(strheader)  
                 current_lines.clear() 
 
+
         if len(current_lines) > 0:
             if len(markdown_tups) == 0:
                 markdown_tups.append((strTitle + strheader , "\n".join(current_lines)))
@@ -62,4 +68,22 @@ class ChunkMarkdownReader(MarkdownReader):
         ]
 
     def _token_size(self, text: str) -> int:
-        return len(self._tokenizer(text))
+        return len(self._tokenizer(text))
+    
+    def findValue(self,expression:str,Field:str):
+        cols = self._colheader.split('|')
+        cols =  [item for item in cols if item]
+
+        for row in self._rows:
+            rowtrs = row.split('|')
+            rowdatas =  [item for item in rowtrs if item and (item!='\r' or item!='\n')]
+            if len(rowdatas) == 0:
+                continue
+            gData = {}
+            for cName,rValue in zip(cols,rowdatas):
+                gData[cName] = rValue
+            if eval(expression,gData):
+                return gData[Field]
+        return ''
+
+

backend/app/engine/loaders/projectJson.py

@@ -55,7 +55,6 @@ class JsonTable:
     def comment(self):
         return self._comment
 
-
 class ProjectJson:
     def __init__(self,dir:str) -> None:
         self._dir = dir
@@ -76,14 +75,5 @@ class ProjectJson:
     def tables(self):
         return self._tables
 
-def getProjectName(dir:str):
-    prjJson = ProjectJson(dir)
-    prjJson.parse()
-    tb:JsonTable = prjJson.table('工程属性')
-    records = tb.records()
-    for record in records:
-        name = record.value('名称') 
-        if name == '工程名称':
-            return record.value('值') 
-    return ''
+