Improve dataset card: Add paper link, project page, task category, and sample usage

README.md

@@ -1,10 +1,97 @@
 ---
-license: cc0-1.0
 language:
 - en
+license: cc0-1.0
 tags:
 - arxiv
+task_categories:
+- text-retrieval
 ---
-Contains the entire arXiv metadata dataset, including 768-dimensional embeddings of all abstracts generated using sentence-transformer/all-MiniLM-L6-v2.
 
-This database is an artifact of the Citegeist project ([Github](https://github.com/chenneking/citegeist))
+# Citegeist: Automated Generation of Related Work Analysis on the arXiv Corpus
+
+This repository contains the entire arXiv metadata dataset, including 768-dimensional embeddings of all abstracts generated using `sentence-transformer/all-MiniLM-L6-v2`. This dataset is a core artifact of the [Citegeist project](https://github.com/chenneking/citegeist).
+
+The dataset and associated project were presented in the paper:
+[Citegeist: Automated Generation of Related Work Analysis on the arXiv Corpus](https://huggingface.co/papers/2503.23229)
+
+Project Page: [https://citegeist.org](https://citegeist.org)
+Code: [https://github.com/chenneking/citegeist](https://github.com/chenneking/citegeist)
+
+## Sample Usage
+
+To generate a related work section for a given abstract, you can use the `citegeist` library's `Generator` class.
+
+First, install the package:
+```bash
+pip install citegeist-arxiv
+```
+
+### Option 1: Using the Hosted Milvus Database
+
+As of March 2025, a hosted Milvus database is provided. You can use it by setting the following environment variables:
+
+```python
+import os
+from citegeist import Generator
+
+os.environ["MILVUS_URI"] = "http://49.12.219.90:19530"
+os.environ["MILVUS_TOKEN"] = "citegeist:citegeist"
+
+# Set your API key for the chosen LLM provider (e.g., for Gemini)
+os.environ["GEMINI_API_KEY"] = "YOUR_GEMINI_API_KEY"
+
+generator = Generator(
+   llm_provider="gemini",  # Choice of: "azure" (OpenAI Studio), "anthropic", "gemini", "mistral", and "openai"
+   api_key=os.environ.get("GEMINI_API_KEY"), # Here, you will need to set the respective API key
+   model_name="gemini-2.0-flash", # Choose the model that the provider supports
+   database_uri=os.environ.get("MILVUS_URI"),  # Set the path (local) / url (remote) for the Milvus DB connection
+   database_token=os.environ.get("MILVUS_TOKEN"),  # Optionally, also set the access token
+)
+# Define input abstract and breadth (5-20), depth (1-5), and diversity (0.0-1.0) parameters.
+abstract = "Large Language Models provide significant new opportunities for the generation of high-quality written works. However, their employment in the research community is inhibited by their tendency to hallucinate invalid sources and lack of direct access to a knowledge base of relevant scientific articles."
+breadth = 10
+depth = 2
+diversity = 0.0
+related_works, citations = generator.generate_related_work(abstract, breadth, depth, diversity)
+
+print("Generated Related Works:\
+", related_works)
+print("\
+Citations:\
+", citations)
+```
+
+### Option 2: Running the Milvus Database Locally
+
+Alternatively, you can run the Milvus database locally. Download [database.db](https://huggingface.co/datasets/chenneking/citegeist-milvus-db/blob/main/database.db) and provide its absolute path as the value for `MILVUS_URI`. You do not need to set `MILVUS_TOKEN` for this local setup.
+
+```python
+import os
+from citegeist import Generator
+
+# Replace <path_to_database.db_goes_here> with the actual absolute path to your database.db file
+os.environ["MILVUS_URI"] = "<path_to_database.db_goes_here>"
+
+# Set your API key for the chosen LLM provider (e.g., for Gemini)
+os.environ["GEMINI_API_KEY"] = "YOUR_GEMINI_API_KEY"
+
+generator = Generator(
+   llm_provider="gemini",  # Choice of: "azure" (OpenAI Studio), "anthropic", "gemini", "mistral", and "openai"
+   api_key=os.environ.get("GEMINI_API_KEY"), # Here, you will need to set the respective API key
+   model_name="gemini-2.0-flash", # Choose the model that the provider supports
+   database_uri=os.environ.get("MILVUS_URI"),  # Set the path (local) database.db file
+)
+# Define input abstract and breadth (5-20), depth (1-5), and diversity (0.0-1.0) parameters.
+abstract = "Large Language Models provide significant new opportunities for the generation of high-quality written works. However, their employment in the research community is inhibited by their tendency to hallucinate invalid sources and lack of direct access to a knowledge base of relevant scientific articles."
+breadth = 10
+depth = 2
+diversity = 0.0
+related_works, citations = generator.generate_related_work(abstract, breadth, depth, diversity)
+
+print("Generated Related Works:\
+", related_works)
+print("\
+Citations:\
+", citations)
+```