check point

2025-12-25 18:06:29 +07:00
commit 31de8b0d84
34 changed files with 3209 additions and 0 deletions
--- a/backup_source/sequence_sentence.py
+++ b/backup_source/sequence_sentence.py
@@ -0,0 +1,127 @@
+"""
+games/sequence_sentence.py - Arrange Sentences Game
+type_id = 2
+"""
+from typing import List
+from pydantic import BaseModel, Field
+from langchain_core.output_parsers import PydanticOutputParser
+
+
+# ============== SCHEMA ==============
+class SentenceItem(BaseModel):
+    sentence: str = Field(description="Full sentence to arrange (EXACT from source)")
+    original_quote: str = Field(description="EXACT quote from source text")
+    image_description: str = Field(default="", description="Visual description of the content")
+    image_keywords: List[str] = Field(default=[], description="Keywords for image search")
+    image_is_complex: bool = Field(default=False, description="True if image needs precise quantities, humans, or multiple detailed objects")
+
+
+class SentenceMetadata(BaseModel):
+    """Metadata đánh giá nội dung"""
+    title: str = Field(
+        description="Title for this content. Prefer title from source document if available and suitable, otherwise create a short descriptive title."
+    )
+    description: str = Field(
+        description="Short description summarizing the content/topic."
+    )
+    grade: int = Field(
+        description="Estimated grade level 1-5 (1=easy/young, 5=advanced/older). Judge by vocabulary, concepts."
+    )
+    type: str = Field(default="sequence_sentence", description="Game type")
+    difficulty: int = Field(
+        description="Difficulty 1-5 for that grade (1=very easy, 5=very hard). Judge by sentence complexity, vocabulary."
+    )
+
+
+class SentenceOutput(BaseModel):
+    """Output wrapper for sentence items"""
+    items: List[SentenceItem] = Field(description="List of sentence items generated from source text")
+    metadata: SentenceMetadata = Field(description="Metadata about the content")
+
+
+# Output parser
+output_parser = PydanticOutputParser(pydantic_object=SentenceOutput)
+
+
+# ============== CONFIG ==============
+GAME_CONFIG = {
+    "game_type": "sequence_sentence",
+    "display_name": "Arrange Sentences",
+    "description": "Arrange sentences in order",
+    "type_id": 2,
+    
+    "active": True,
+    
+    "max_items": 10,
+    "schema": SentenceItem,
+    "output_schema": SentenceOutput,
+    "output_parser": output_parser,
+    
+    # Dùng cho analyze + generate (không có format rules)
+    "system_prompt": """Extract sentences from source text.
+
+RULES:
+1. KEEP THE ORIGINAL LANGUAGE - Do NOT translate
+2. sentence = EXACT copy from source text
+3. original_quote = same as sentence value
+4. image_description = ALWAYS provide a short visual description (NEVER empty)
+5. image_is_complex = FALSE for simple/static objects, TRUE for quantities/humans/complex scenes""",
+    
+    # Dùng cho generate trực tiếp (CÓ format rules)
+    "direct_prompt": """Extract sentences from source text.
+
+EXPECTED INPUT: List of sentences (separated by semicolon, newline, or similar)
+
+STEP 1 - VALIDATE INPUT:
+Analyze if input looks like a list of sentences suitable for "arrange sentences" game.
+- Should contain multiple complete sentences
+- Should NOT be a quiz, single word list, or Q&A format
+
+If input is clearly NOT suitable (e.g. it's a quiz, single words only, or wrong format), return:
+{{"items": [], "format_error": "Input không phù hợp cho game sắp xếp câu"}}
+
+STEP 2 - EXTRACT (if valid):
+RULES:
+1. KEEP THE ORIGINAL LANGUAGE - Do NOT translate
+2. Extract ALL sentences from source
+3. sentence = EXACT sentence from source (trim whitespace)
+4. original_quote = same as sentence value
+5. image_description = ALWAYS provide a short visual description (NEVER leave empty)
+6. image_is_complex:
+   - FALSE: simple objects, static things, no specific quantities (e.g. "a sun", "a tree")
+   - TRUE: needs exact quantities, humans/people, or complex details (e.g. "3 birds", "a boy reading")""",
+}
+
+
+# ============== EXAMPLES ==============
+EXAMPLES = [
+    {
+        "input": "The Sun is a star; The Moon orbits Earth; Mars is red",
+        "output": {
+            "items": [
+                {
+                    "sentence": "The Sun is a star",
+                    "original_quote": "The Sun is a star",
+                    "image_description": "A bright glowing sun",
+                    "image_keywords": ["sun", "star"],
+                    "image_is_complex": False
+                },
+                {
+                    "sentence": "The Moon orbits Earth",
+                    "original_quote": "The Moon orbits Earth",
+                    "image_description": "Moon circling around Earth",
+                    "image_keywords": ["moon", "earth", "orbit"],
+                    "image_is_complex": False
+                },
+                {
+                    "sentence": "Mars is red",
+                    "original_quote": "Mars is red",
+                    "image_description": "Red planet Mars",
+                    "image_keywords": ["mars", "red", "planet"],
+                    "image_is_complex": False
+                }
+            ]
+        },
+        "why_suitable": "Source has sentences separated by semicolons"
+    }
+]