'commit'

4 weeks ago · 9998fa723c
parent a1abe0fe3e
commit 9998fa723c
4 changed files with 48 additions and 6 deletions
--- a/dsRag/Doc/10、下载NLTK库.md
+++ b/dsRag/Doc/10、下载NLTK库.md
@ -0,0 +1,5 @@
+### 手动下载nltk库
+https://blog.csdn.net/QM19900420/article/details/128100591
+
+### Python NLTK库【NLP核心库】全面解析
+https://blog.csdn.net/webcai_3/article/details/147347144
--- a/dsRag/Start.py
+++ b/dsRag/Start.py
@ -121,6 +121,13 @@ async def generate_stream(client, milvus_pool, collection_manager, query, docume
                            # 获取完整内容
                            full_content = record['tags'].get('full_content', record['user_input'])
                            context = context + full_content
+                        else:
+                            logger.warning(f"距离太远，忽略此结果: {hit.id}")
+                            logger.info(f"标签: {record['tags']}")
+                            logger.info(f"用户问题: {record['user_input']}")
+                            logger.info(f"时间: {record['timestamp']}")
+                            logger.info(f"距离: {hit.distance}")
+                            continue

                    except Exception as e:
                        logger.error(f"查询失败: {e}")
--- a/dsRag/Test/expand_with_synonyms.py
+++ b/dsRag/Test/expand_with_synonyms.py
@ -0,0 +1,21 @@
+from nltk.corpus import wordnet
+import jieba
+
+def expand_with_synonyms(query):
+    words = jieba.lcut(query)
+    expanded = []
+    for word in words:
+        synonyms = set()
+        for syn in wordnet.synsets(word, lang='cmn'):
+            for lemma in syn.lemma_names('cmn'):
+                synonyms.add(lemma)
+        if synonyms:
+            expanded.append(f"({'|'.join(synonyms)})")
+        else:
+            expanded.append(word)
+    return ' '.join(expanded)
+
+original_query = "微积分的基本定理是什么？"
+expanded_query = expand_with_synonyms(original_query)
+print(f"原始查询: {original_query}")
+print(f"扩展后查询: {expanded_query}")
--- a/dsRag/static/ai.html
+++ b/dsRag/static/ai.html
@ -3,7 +3,7 @@
 <head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <title>小学数学问答系统</title>
+    <title>问答系统</title>
    <style>
        body {
            font-family: 'Microsoft YaHei', sans-serif;
@ -182,7 +182,7 @@
 </head>
 <body>
 <div class="container">
-    <h1>【小学数学】大模型知识库</h1>
+    <h1>【东师理想】教学大模型知识库</h1>
    <div class="data-area" id="answerArea">
        <div style="color:#666; padding:20px; text-align:center;">
            <p>请在下方输入您的问题，答案将在此处显示</p>
@ -201,6 +201,7 @@
                    <div class="example-item" onclick="fillExample('帮我写一下 “如何理解点、线、面、体、角”的教学设计')">
                        帮我写一下 “如何理解点、线、面、体、角”的教学设计
                    </div>
+                    <div class="example-item" onclick="fillExample('文言虚词')">请解释一下“文言虚词”</div>
                </div>
            </div>

@ -215,20 +216,28 @@
        </div>

        <div class="doc-links">
-            <h3>知识库：</h3>
+            <h3>知识库范围</h3>
            <div class="doc-checkboxes">
                <label>
-                    <input type="checkbox" name="documents" checked value="MATH_DATA_1">
+                    <input type="checkbox" name="documents" value="MATH_DATA_1">
                    小学数学教学中的若干问题
                </label>
                <label>
                    <input type="checkbox" name="documents" value="MATH_DATA_2">
                    小学数学知识点
                </label>
+                <label>
+                    <input type="checkbox" name="documents" value="CHINESE_DATA_1">
+                    高中语文文言文
+                </label>
            </div>
-            <a href="/static/Txt/小学数学教学中的若干问题.docx" class="doc-link">下载</a>
-            <a href="/static/Txt/小学数学知识点.docx" class="doc-link">下载</a>
        </div>
+        <div><h3>知识库文稿下载</h3>
+            <a href="/static/Txt/小学数学教学中的若干问题.docx" class="doc-link">小学数学教学中的若干问题</a>
+            <a href="/static/Txt/小学数学知识点.docx" class="doc-link">小学数学知识点</a>
+            <a href="/static/Txt/高中文言文.docx" class="doc-link">高中语文文言文</a>
+        </div>
+        <br>

        <button id="submitBtn" onclick="submitQuestion()"><span class="icon">💡</span>提问</button>
        <button id="clearBtn" onclick="clearAll()"><span class="icon">🗑️</span>清空</button>