fix(llm): definitively resolve NameError in prompt creation

.gitignore

@@ -1,2 +1,3 @@
 venv
 __pycache__
+academic_article.md

pipeline/llm_service.py

@@ -406,15 +406,9 @@ class LLMService:
 
 ## Introduction
 
-Below is a table of similarity metrics for a collection of Tibetan texts. Your task is to provide a scholarly analysis of these results. Focus on:
-
-1.  **Overall Similarity**: Are the texts generally similar or dissimilar?
-2.  **Key Relationships**: Which pairs of texts are most and least similar? What might this imply about their relationship (e.g., different recensions, direct copies, shared sources)?
-3.  **Metric-Specific Insights**: What does each metric (Jaccard, LCS, TF-IDF, Semantic) reveal individually?
-4.  **Synthesis**: Combine the insights from all metrics to form a comprehensive conclusion about the textual relationships.
-
-## Data
+You will be provided with a table of text similarity scores in Markdown format. Your task is to provide a scholarly interpretation of these results for an academic article on Tibetan textual analysis. Do not simply restate the data. Instead, focus on the *implications* of the scores. What do they suggest about the relationships between the texts? Consider potential reasons for both high and low similarity across different metrics (e.g., shared vocabulary vs. structural differences).
 
+**Data:**
 {md_table}
 
 ## Instructions
@@ -422,8 +416,7 @@ Below is a table of similarity metrics for a collection of Tibetan texts. Your t
 Your analysis will be performed using the `{model_name}` model. Provide a concise, scholarly analysis in well-structured markdown.
 """
         
-        # Replace [CSV_DATA] with the actual CSV data
-        prompt = prompt.replace("[CSV_DATA]", csv_data)
+
         
         return prompt