add_subcategories

app.py

@@ -3,21 +3,51 @@ from google import genai
 from google.genai import types
 import os
 import requests
+import json
 
 # setting api keys
 genai_api_key = os.environ.get('GENAI_API_KEY')
 client = genai.Client(api_key=genai_api_key)
 
-# reading arxiv category list
-with open("./arxiv_categories.txt", "r") as f:
-    cat_list = [line for line in f]
+# Load categories from JSON file
+with open("./arxiv_categories.json", "r") as f:
+    arxiv_categories = json.load(f)
 
-def launch(temp, cat, iteration):
+# Extract main categories and sub-categories
+cat_list = list(arxiv_categories.keys())
+sub_categories = {
+    cat: list(arxiv_categories[cat]["subcategories"].keys())
+    for cat in cat_list
+}
+
+# Create a mapping of sub-categories to their descriptions
+sub_category_descriptions = {
+    cat: {
+        sub: arxiv_categories[cat]["subcategories"][sub]
+        for sub in arxiv_categories[cat]["subcategories"]
+    }
+    for cat in cat_list
+}
+
+def launch(temp, cat, sub_cat, iteration):
     yield "⏳ Generating summaries... please wait."
 
+    # Extract the category code from the full string (e.g., "astro-ph - Astrophysics" -> "astro-ph")
+    category_code = cat.split(" - ")[0]
+    
+    # Extract sub-category codes from the full strings
+    sub_category_codes = [sub.split(" - ")[0] for sub in sub_cat] if sub_cat else []
+
+    # Construct the category string for arXiv API
+    if sub_category_codes:  # If sub-categories are selected
+        # Format: main_category.sub_category
+        categories = [f"{category_code}.{sub}" for sub in sub_category_codes]
+        cat_str = " OR ".join(categories)
+    else:  # If no sub-categories selected, use main category
+        cat_str = category_code
+
     # arxiv announcement page
-    cat_txt = cat.split()
-    url = f"https://arxiv.org/list/{cat_txt[-1][1:-1]}/new"
+    url = f"https://arxiv.org/list/{cat_str}/new"
     url = "https://r.jina.ai/" + url
     scraped_text = requests.get(url)
 
@@ -26,12 +56,13 @@ def launch(temp, cat, iteration):
     ind2 = scraped_text.text.index("Cross submissions (")
     filtered = scraped_text.text[ind1:ind2]
 
-
-    prompt = f"""You are an expert in { " ".join(cat_txt[:-1]) } and a scientific assistant.
+    # Update the prompt to include the selected categories
+    category_description = f"{category_code}" + (f" ({', '.join(sub_category_codes)})" if sub_category_codes else "")
+    prompt = f"""You are an expert in {category_description} and a scientific assistant.
 You will be given a list of new paper entries from the arXiv daily announcement page.
 
 Each entry includes the title and abstract. 
-For each entry, extract insights to help readers understand the paper’s purpose, methods, and findings. 
+For each entry, extract insights to help readers understand the paper's purpose, methods, and findings. 
 If any information is not clearly mentioned in the abstract, write `"Not specified."` 
 
 For each paper, briefly answer:
@@ -93,8 +124,6 @@ Please do NOT generate any HTML, only return answers in plain structured text.
         if 'Article link:' in paper:
             article_link = paper.split('Article link:')[1].strip()
 
-
-        
         html_output += f"""
         <div style="border:1px solid #ddd; padding:20px; margin:15px; border-radius:15px; background: linear-gradient(to right, #f7f7f7, #e1e1e1); box-shadow: 0px 8px 15px rgba(0,0,0,0.1); transition: transform 0.3s ease-in-out, box-shadow 0.3s ease-in-out; font-family: 'Verdana', sans-serif; color: #34495e;">
             <h3 style="font-size: 22px; font-weight: bold; margin-bottom: 12px; color: #2c3e50; text-shadow: 1px 1px 3px rgba(0,0,0,0.2);">{[idx+1]} {title[:-2]}</h3>
@@ -130,15 +159,71 @@ css = """
 """
 # font-family: Arial, sans-serif; /* Apply Arial font to all text */
 with gr.Blocks(theme=gr.themes.Glass(), css=css) as demo:
-    gr.Markdown("## Hey, I’m skimarXiv 👋  I summarize today’s arXiv papers from your favorite fields, saving you time so you don't have to!")
+    gr.Markdown("## Hey, I'm skimarXiv 👋  I summarize today's arXiv papers from your favorite fields, saving you time so you don't have to!")
     gr.Markdown("SkimArXiv provides daily AI-generated summaries of new arXiv papers (using Gemini) — customized by the research fields you select. Stay up-to-date with the latest discoveries across science, technology, math, and more.")
+    
     with gr.Row():
         slider_temp = gr.Slider(minimum=0, maximum=2, value=0.95, label="temperature", info="0 for reproducibility and > 0 for creativity. 0.95 is the default value for gemini-2.0-flash.")
         iterations = gr.Number(value=2, label="iterations", info="larger number of papers requires larger iterations (e.g., 1 iteration ~ 40-50 papers).")
-        category = gr.Dropdown(cat_list, label="arXiv category", info="All subjects are included from Physics to Economics.")
+        
+        with gr.Column():
+            # Set default category to astro-ph and initialize its sub-categories
+            default_category = 'astro-ph'
+            
+            # Create a mapping of category codes to their descriptions for the dropdown
+            category_choices = {
+                cat: f"{cat} - {arxiv_categories[cat]['description']}"
+                for cat in cat_list
+            }
+            
+            category = gr.Dropdown(
+                choices=list(category_choices.values()),
+                value=category_choices[default_category],
+                label="arXiv category",
+                info="Select a main category"
+            )
+            
+            # Create a mapping of sub-category codes to their descriptions
+            def get_sub_category_choices(category_code):
+                if category_code in sub_categories and sub_categories[category_code]:
+                    return {
+                        sub: f"{sub} - {sub_category_descriptions[category_code][sub]}"
+                        for sub in sub_categories[category_code]
+                    }
+                return {}
+            
+            sub_category = gr.CheckboxGroup(
+                choices=list(get_sub_category_choices(default_category).values()),
+                value=[],
+                label="Sub-categories",
+                info="Select specific sub-categories (optional)",
+                interactive=True,
+                visible=bool(sub_categories[default_category])
+            )
+            
+            # Add change event to update sub-categories when main category changes
+            def update_sub_categories(category_full):
+                # Extract the category code from the full string (e.g., "astro-ph - Astrophysics" -> "astro-ph")
+                category_code = category_full.split(" - ")[0]
+                
+                if category_code in sub_categories and sub_categories[category_code]:
+                    sub_choices = get_sub_category_choices(category_code)
+                    return {
+                        sub_category: gr.CheckboxGroup(
+                            choices=list(sub_choices.values()),
+                            value=[],
+                            visible=True
+                        )
+                    }
+                return {
+                    sub_category: gr.CheckboxGroup(choices=[], value=[], visible=False)
+                }
+            
+            category.change(fn=update_sub_categories, inputs=category, outputs=[sub_category])
+        
         generate_button = gr.Button("Generate Today's Summaries")
 
     output_html = gr.HTML()
-    generate_button.click(fn=launch, inputs=[slider_temp, category, iterations], outputs=output_html)
+    generate_button.click(fn=launch, inputs=[slider_temp, category, sub_category, iterations], outputs=output_html)
 
-demo.launch()
+demo.launch(share=True)

arxiv_categories.json

@@ -0,0 +1,238 @@
+{
+  "astro-ph": {
+    "description": "Astrophysics",
+    "subcategories": {
+      "GA": "Astrophysics of Galaxies",
+      "CO": "Cosmology and Nongalactic Astrophysics",
+      "EP": "Earth and Planetary Astrophysics",
+      "IM": "Instrumentation and Methods for Astrophysics",
+      "SR": "Solar and Stellar Astrophysics",
+      "HE": "High Energy Astrophysical Phenomena"
+    }
+  },
+  "cond-mat": {
+    "description": "Condensed Matter",
+    "subcategories": {
+      "dis-nn": "Disordered Systems and Neural Networks",
+      "mes-hall": "Mesoscale and Nanoscale Physics",
+      "mtrl-sci": "Materials Science",
+      "other": "Other Condensed Matter",
+      "quant-gas": "Quantum Gases",
+      "soft": "Soft Condensed Matter",
+      "stat-mech": "Statistical Mechanics",
+      "str-el": "Strongly Correlated Electrons",
+      "supr-con": "Superconductivity"
+    }
+  },
+  "gr-qc": {
+    "description": "General Relativity and Quantum Cosmology",
+    "subcategories": {}
+  },
+  "hep-ex": {
+    "description": "High Energy Physics - Experiment",
+    "subcategories": {}
+  },
+  "hep-lat": {
+    "description": "High Energy Physics - Lattice",
+    "subcategories": {}
+  },
+  "hep-ph": {
+    "description": "High Energy Physics - Phenomenology",
+    "subcategories": {}
+  },
+  "hep-th": {
+    "description": "High Energy Physics - Theory",
+    "subcategories": {}
+  },
+  "math-ph": {
+    "description": "Mathematical Physics",
+    "subcategories": {}
+  },
+  "nlin": {
+    "description": "Nonlinear Sciences",
+    "subcategories": {
+      "AO": "Adaptation and Self-Organizing Systems",
+      "CD": "Chaotic Dynamics",
+      "CG": "Cellular Automata and Lattice Gases",
+      "PS": "Pattern Formation and Solitons",
+      "SI": "Exactly Solvable and Integrable Systems"
+    }
+  },
+  "nucl-ex": {
+    "description": "Nuclear Experiment",
+    "subcategories": {}
+  },
+  "nucl-th": {
+    "description": "Nuclear Theory",
+    "subcategories": {}
+  },
+  "physics": {
+    "description": "Physics",
+    "subcategories": {
+      "acc-ph": "Accelerator Physics",
+      "app-ph": "Applied Physics",
+      "ao-ph": "Atmospheric and Oceanic Physics",
+      "atom-ph": "Atomic Physics",
+      "atm-clus": "Atomic and Molecular Clusters",
+      "bio-ph": "Biological Physics",
+      "chem-ph": "Chemical Physics",
+      "class-ph": "Classical Physics",
+      "comp-ph": "Computational Physics",
+      "data-an": "Data Analysis, Statistics and Probability",
+      "flu-dyn": "Fluid Dynamics",
+      "gen-ph": "General Physics",
+      "geo-ph": "Geophysics",
+      "hist-ph": "History and Philosophy of Physics",
+      "ins-det": "Instrumentation and Detectors",
+      "med-ph": "Medical Physics",
+      "optics": "Optics",
+      "plasm-ph": "Plasma Physics",
+      "pop-ph": "Popular Physics",
+      "soc-ph": "Physics and Society",
+      "space-ph": "Space Physics"
+    }
+  },
+  "quant-ph": {
+    "description": "Quantum Physics",
+    "subcategories": {}
+  },
+  "math": {
+    "description": "Mathematics",
+    "subcategories": {
+      "AC": "Commutative Algebra",
+      "AG": "Algebraic Geometry",
+      "AP": "Analysis of PDEs",
+      "AT": "Algebraic Topology",
+      "CA": "Classical Analysis and ODEs",
+      "CO": "Combinatorics",
+      "CT": "Category Theory",
+      "CV": "Complex Variables",
+      "DG": "Differential Geometry",
+      "DS": "Dynamical Systems",
+      "FA": "Functional Analysis",
+      "GM": "General Mathematics",
+      "GN": "General Topology",
+      "GR": "Group Theory",
+      "GT": "Geometric Topology",
+      "HO": "History and Overview",
+      "IT": "Information Theory",
+      "KT": "K-Theory and Homology",
+      "LO": "Logic",
+      "MG": "Metric Geometry",
+      "MP": "Mathematical Physics",
+      "NA": "Numerical Analysis",
+      "NT": "Number Theory",
+      "OA": "Operator Algebras",
+      "OC": "Optimization and Control",
+      "PR": "Probability",
+      "QA": "Quantum Algebra",
+      "RA": "Rings and Algebras",
+      "RT": "Representation Theory",
+      "SG": "Symplectic Geometry",
+      "SP": "Spectral Theory",
+      "ST": "Statistics Theory"
+    }
+  },
+  "cs": {
+    "description": "Computing Research Repository",
+    "subcategories": {
+      "AI": "Artificial Intelligence",
+      "AR": "Hardware Architecture",
+      "CC": "Computational Complexity",
+      "CE": "Computational Engineering, Finance, and Science",
+      "CG": "Computational Geometry",
+      "CL": "Computation and Language",
+      "CR": "Cryptography and Security",
+      "CV": "Computer Vision and Pattern Recognition",
+      "CY": "Computers and Society",
+      "DB": "Databases",
+      "DC": "Distributed, Parallel, and Cluster Computing",
+      "DL": "Digital Libraries",
+      "DM": "Discrete Mathematics",
+      "DS": "Data Structures and Algorithms",
+      "ET": "Emerging Technologies",
+      "FL": "Formal Languages and Automata Theory",
+      "GL": "General Literature",
+      "GR": "Graphics",
+      "GT": "Computer Science and Game Theory",
+      "HC": "Human-Computer Interaction",
+      "IR": "Information Retrieval",
+      "IT": "Information Theory",
+      "LG": "Machine Learning",
+      "LO": "Logic in Computer Science",
+      "MA": "Multiagent Systems",
+      "MM": "Multimedia",
+      "MS": "Mathematical Software",
+      "NA": "Numerical Analysis",
+      "NE": "Neural and Evolutionary Computing",
+      "NI": "Networking and Internet Architecture",
+      "OH": "Other Computer Science",
+      "OS": "Operating Systems",
+      "PF": "Performance",
+      "PL": "Programming Languages",
+      "RO": "Robotics",
+      "SC": "Symbolic Computation",
+      "SD": "Sound",
+      "SE": "Software Engineering",
+      "SI": "Social and Information Networks",
+      "SY": "Systems and Control"
+    }
+  },
+  "q-bio": {
+    "description": "Quantitative Biology",
+    "subcategories": {
+      "BM": "Biomolecules",
+      "CB": "Cell Behavior",
+      "GN": "Genomics",
+      "MN": "Molecular Networks",
+      "NC": "Neurons and Cognition",
+      "OT": "Other Quantitative Biology",
+      "PE": "Populations and Evolution",
+      "QM": "Quantitative Methods",
+      "SC": "Subcellular Processes",
+      "TO": "Tissues and Organs"
+    }
+  },
+  "q-fin": {
+    "description": "Quantitative Finance",
+    "subcategories": {
+      "CP": "Computational Finance",
+      "EC": "Economics",
+      "GN": "General Finance",
+      "MF": "Mathematical Finance",
+      "PM": "Portfolio Management",
+      "PR": "Pricing of Securities",
+      "RM": "Risk Management",
+      "ST": "Statistical Finance",
+      "TR": "Trading and Market Microstructure"
+    }
+  },
+  "stat": {
+    "description": "Statistics",
+    "subcategories": {
+      "AP": "Applications",
+      "CO": "Computation",
+      "ME": "Methodology",
+      "ML": "Machine Learning",
+      "OT": "Other Statistics",
+      "TH": "Statistics Theory"
+    }
+  },
+  "eess": {
+    "description": "Electrical Engineering and Systems Science",
+    "subcategories": {
+      "AS": "Audio and Speech Processing",
+      "IV": "Image and Video Processing",
+      "SP": "Signal Processing",
+      "SY": "Systems and Control"
+    }
+  },
+  "econ": {
+    "description": "Economics",
+    "subcategories": {
+      "EM": "Econometrics",
+      "GN": "General Economics",
+      "TH": "Theoretical Economics"
+    }
+  }
+}