.

app.py

@@ -3,7 +3,6 @@ import gradio as gr
 from init import get_secrets, initialize_data, update_dataframe
 from gen.openllm import GradioLLaMA2ChatPPManager, GradioMistralChatPPManager
 from gen.gemini_chat import GradioGeminiChatPPManager
-from constants.styles import STYLE
 from constants.js import (
     UPDATE_SEARCH_RESULTS, OPEN_CHAT_IF,
     CLOSE_CHAT_IF, UPDATE_CHAT_HISTORY
@@ -31,7 +30,7 @@ last_day = sorted_day[-1]
 last_papers = date_dict[last_year][last_month][last_day]
 selected_paper = last_papers[0]
 
-with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
+with gr.Blocks(css="constants/styles.css", theme=gr.themes.Soft()) as demo:
     cur_arxiv_id = gr.Textbox(selected_paper['arxiv_id'], visible=False)
     local_data = gr.JSON({}, visible=False)
     chat_state = gr.State({
@@ -88,14 +87,11 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
         # with gr.Row():
         with gr.Row():
             arxiv_link = gr.Markdown(
-                "[![arXiv](https://img.shields.io/badge/arXiv-%s-b31b1b.svg?style=for-the-badge)](https://arxiv.org/abs/%s)" % (selected_paper['arxiv_id'], selected_paper['arxiv_id']),
-                elem_classes=["markdown-center"]
+                "[![arXiv](https://img.shields.io/badge/arXiv-%s-b31b1b.svg?style=for-the-badge)](https://arxiv.org/abs/%s)" % (selected_paper['arxiv_id'], selected_paper['arxiv_id']) + " "
+                "[![Paper page](https://huggingface.co/datasets/huggingface/badges/resolve/main/paper-page-lg.svg)](https://huggingface.co/papers/%s)" % selected_paper['arxiv_id'] + " ",
+                elem_id="link-md",
             )
-            hf_paper_link = gr.Markdown(
-                "[![Paper page](https://huggingface.co/datasets/huggingface/badges/resolve/main/paper-page-lg.svg)](https://huggingface.co/papers/%s)" % selected_paper['arxiv_id'],
-                elem_classes=["markdown-center"]
-            )
-            chat_button = gr.Button("💬 about paper", interactive=True, elem_id="chat-button")
+        chat_button = gr.Button("Chat about any custom questions", interactive=True, elem_id="chat-button")
             
         summary = gr.Markdown(f"{selected_paper['summary']}", elem_classes=["small-font"])
 
@@ -154,23 +150,21 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
                     breath_q_eli5_2 = gr.Markdown(f"↪ **(ELI5)** {selected_paper['2_additional_breath_q:answers:eli5']}", elem_classes=["small-font"])
                     breath_q_expert_2 = gr.Markdown(f"↪ **(Technical)** {selected_paper['2_additional_breath_q:answers:expert']}", visible=False, elem_classes=["small-font"])
 
-        with gr.Column(visible=True):
-            gr.Markdown("## Request any arXiv ids")
-            arxiv_queue = gr.Dataframe(
-                headers=["Requested arXiv IDs"], col_count=(1, "fixed"),
-                value=requested_arxiv_ids_df,
-                datatype=["str"],
-                interactive=False,
-            )
-    
-            arxiv_id_enter = gr.Textbox(placeholder="Enter comma separated arXiv IDs...", elem_classes=["textbox-no-label"])
-            arxiv_id_enter.submit(
-                add_arxiv_ids_to_queue,
-                [arxiv_queue, arxiv_id_enter],
-                [arxiv_queue, arxiv_id_enter],
-                concurrency_limit=20,
-            )
-
+        gr.Markdown("## Request any arXiv ids")
+        arxiv_queue = gr.Dataframe(
+            headers=["Requested arXiv IDs"], col_count=(1, "fixed"),
+            value=requested_arxiv_ids_df,
+            datatype=["str"],
+            interactive=False,
+        )
+
+        arxiv_id_enter = gr.Textbox(placeholder="Enter comma separated arXiv IDs...", elem_classes=["textbox-no-label"])
+        arxiv_id_enter.submit(
+            add_arxiv_ids_to_queue,
+            [arxiv_queue, arxiv_id_enter],
+            [arxiv_queue, arxiv_id_enter],
+            concurrency_limit=20,
+        )
 
     gr.Markdown("The target papers are collected from [Hugging Face 🤗 Daily Papers](https://huggingface.co/papers) on a daily basis. "
                 "The entire data is generated by [Google's Gemini 1.0](https://deepmind.google/technologies/gemini/) Pro. "
@@ -251,7 +245,7 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
         set_paper, [year_dd, month_dd, day_dd, papers_dd],
         [
             cur_arxiv_id,
-            title, arxiv_link, hf_paper_link, summary,
+            title, arxiv_link, summary,
             basic_q_0, basic_q_eli5_0, basic_q_expert_0,
             depth_q_0, depth_q_eli5_0, depth_q_expert_0,
             breath_q_0, breath_q_eli5_0, breath_q_expert_0,
@@ -271,7 +265,7 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
         set_paper, [year_dd, month_dd, day_dd, papers_dd],
         [
             cur_arxiv_id,
-            title, arxiv_link, hf_paper_link, summary,
+            title, arxiv_link, summary,
             basic_q_0, basic_q_eli5_0, basic_q_expert_0,
             depth_q_0, depth_q_eli5_0, depth_q_expert_0,
             breath_q_0, breath_q_eli5_0, breath_q_expert_0,
@@ -284,14 +278,14 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
             depth_q_2, depth_q_eli5_2, depth_q_expert_2,
             breath_q_2, breath_q_eli5_2, breath_q_expert_2
         ],
-        concurrency_limit=20,
+        concurrency_limit=20, 
     )
 
     day_dd.input(get_paper_by_day, inputs=[year_dd, month_dd, day_dd], outputs=[papers_dd]).then(
         set_paper, [year_dd, month_dd, day_dd, papers_dd],
         [
             cur_arxiv_id,
-            title, arxiv_link, hf_paper_link, summary,
+            title, arxiv_link, summary,
             basic_q_0, basic_q_eli5_0, basic_q_expert_0,
             depth_q_0, depth_q_eli5_0, depth_q_expert_0,
             breath_q_0, breath_q_eli5_0, breath_q_expert_0,
@@ -310,7 +304,7 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
     papers_dd.change(set_paper, [year_dd, month_dd, day_dd, papers_dd],
         [
             cur_arxiv_id,
-            title, arxiv_link, hf_paper_link, summary,
+            title, arxiv_link, summary,
             basic_q_0, basic_q_eli5_0, basic_q_expert_0,
             depth_q_0, depth_q_eli5_0, depth_q_expert_0,
             breath_q_0, breath_q_eli5_0, breath_q_expert_0,
@@ -343,7 +337,8 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
             basic_q_eli5_0, basic_q_expert_0, depth_q_eli5_0, depth_q_expert_0, breath_q_eli5_0, breath_q_expert_0,
             basic_q_eli5_1, basic_q_expert_1, depth_q_eli5_1, depth_q_expert_1, breath_q_eli5_1, breath_q_expert_1,
             basic_q_eli5_2, basic_q_expert_2, depth_q_eli5_2, depth_q_expert_2, breath_q_eli5_2, breath_q_expert_2
-        ]
+        ],
+        concurrency_limit=20,
     )
     
     chat_button.click(None, [cur_arxiv_id], [local_data, chatbot], js=OPEN_CHAT_IF)
@@ -375,7 +370,7 @@ with gr.Blocks(css=STYLE, theme=gr.themes.Soft()) as demo:
         js=UPDATE_CHAT_HISTORY
     )
 
-    demo.load(lambda: update_dataframe(request_arxiv_repo_id), None, arxiv_queue, every=360)
+    demo.load(lambda: update_dataframe(request_arxiv_repo_id), None, arxiv_queue, every=180)
     # demo.load(None, None, [chatbot, local_data], js=GET_LOCAL_STORAGE % idx.value)
 
 start_date = datetime.now() + timedelta(minutes=1)
@@ -399,5 +394,5 @@ demo.queue(
     default_concurrency_limit=20,
     max_size=256
 ).launch(
-    debug=True
+    share=True, debug=True
 )

background.py

@@ -18,6 +18,17 @@ def _filter_function(example, ids):
     print(example)
     return example
 
+def _extract_arxiv_id(text):
+    print(text)
+    start = text.find("[") + 1
+    end = text.find("]")
+
+    # Extract the text between brackets
+    if start != -1 and end != -1:
+        return text[start:end]
+    else:
+        return text
+
 def process_arxiv_ids(gemini_api, hf_repo_id, req_hf_repo_id, hf_token, restart_repo_id, how_many=10):
     arxiv_ids = []
 
@@ -25,7 +36,7 @@ def process_arxiv_ids(gemini_api, hf_repo_id, req_hf_repo_id, hf_token, restart_
     for d in ds1['train']:
         req_arxiv_ids = d['Requested arXiv IDs']
         if len(req_arxiv_ids) > 0 and req_arxiv_ids[0] != "top":
-            arxiv_ids = arxiv_ids + req_arxiv_ids
+            arxiv_ids = arxiv_ids + [_extract_arxiv_id(req_arxiv_ids[0])]
 
     arxiv_ids = arxiv_ids[:how_many]
 
@@ -77,7 +88,7 @@ def process_arxiv_ids(gemini_api, hf_repo_id, req_hf_repo_id, hf_token, restart_
 
                 print(f"......Updating request arXiv HF Dataset repo at [{req_hf_repo_id}]")
                 ds1 = ds1['train'].map(
-                    lambda example: _filter_function(example, [arxiv_id])
+                    lambda example: _filter_function(example, [f"[{arxiv_id}] {title}"])
                 ).filter(
                     lambda example: len(example['Requested arXiv IDs']) > 0
                 )

constants/styles.css

@@ -0,0 +1,199 @@
+.main {
+    @media only screen and (min-width: 1000px) { 
+        width: 70% !important;
+        margin: 0 auto; /* Center the container */
+    }
+}
+
+.small-font{
+  font-size: 12pt !important;
+  transition: font-size 0.4s ease-out; 
+  transition-delay: 0.8s;
+}
+
+.small-font:hover {
+  font-size: 20px !important;
+  transition: font-size 0.3s ease-out;
+  transition-delay: 1.5s;
+}
+
+.group {
+  padding-top: 10px;
+  padding-left: 10px;
+  padding-right: 10px;
+  padding-bottom: 10px;
+  border: 2px dashed gray;
+  border-radius: 20px;
+  box-shadow: 5px 3px 10px 1px rgba(0, 0, 0, 0.4) !important;
+}
+
+.accordion > button > span{
+  font-size: 12pt !important;
+}
+
+.accordion {
+  border-style: dashed !important;
+  border-left-width: 2px !important;
+  border-bottom-width: 2.5px !important;
+  border-top: none !important;
+  border-right: none !important;
+  box-shadow: none !important;
+}
+
+.no-gap {
+    gap: 0px;
+}
+
+.no-radius {
+    border-radius: 0px;    
+}
+
+.textbox-no-label > label > span {
+    display: none;
+}
+
+.exp-type > span {
+    display: none;
+}
+
+.conv-type > span {
+    display: none;
+}
+
+.conv-type .wrap:nth-child(3) {
+    width: 167px;
+    margin: auto;    
+}
+
+button {
+    font-size: 10pt !important;
+}
+
+h3 {
+    font-size: 13pt !important;
+}
+
+#control-panel {
+    margin-bottom: 30px;
+}
+
+#chatbot {
+    background-color: white;
+    border: 1px solid #ccc;
+    padding: 20px;
+    box-shadow: 0px 5px 5px rgba(0, 0, 0, 0.3);
+    border-radius: 30px;
+    height: 80%;
+    width: 80%;
+
+    position: fixed;
+    top: 50%;
+    left: 50%;
+    transform: translate(-50%, -50%);
+    z-index: 1000; /* Or a high enough value to stay on top */    
+
+    @media (max-width: 768px) { /* Adjust this breakpoint as needed */
+        width: 95%; 
+    }
+
+    @media (prefers-color-scheme: dark) {
+        background-color: dimgrey;
+    }    
+}
+
+#chat-button {
+    border-radius: 30px;
+    font-size: 13pt !important;
+
+    @media only screen and (min-width: 500px) { 
+        font-size: 10pt;
+        margin: auto; /* Center the container */
+    }
+}
+
+#chatbot-inside {
+    height: 100% !important;
+    border-width: 1px !important;
+    border-color: lightgray !important;
+}
+
+#chatbot-txtbox {
+    padding-bottom: 25px;
+}
+
+#chatbot-bottm {
+    padding-left: 10px;
+    padding-right: 10px;
+}
+
+#chatbot-right-button {
+    float: right;
+    width: 20px;
+    font-size: 17pt;
+}
+
+#chatbot-info {
+    word-break: break-word;
+}
+
+#chatbot-back {
+    position: absolute; /* Stay in place even when scrolling */
+    z-index: 1000; /* Ensure it's on top of everything else */
+    width: 100%;
+    height: 100%;
+    left: 0px;
+    top: 0px;
+
+    opacity: 0;
+    visibility: hidden; /* Ensures the element is not interactive */
+    transition: opacity 0.5s ease, visibility 0s 0.5s; /* Transition for opacity and delay visibility */
+}
+
+#chatbot-back.visible {
+    opacity: 1;
+    visibility: visible; /* Now visible and interactive */
+    transition: opacity 0.5s ease; /* Smooth transition for opacity */
+}
+
+.hover-opacity {
+  opacity: 0.8;  /* Normal opacity of the element */
+  transition: opacity 0.3s ease-in-out;  /* Smooth opacity change */
+}
+
+.hover-opacity:hover {
+  opacity: 1;  /* Full opacity on hover */
+}
+
+.markdown-center {
+    display: block;
+    text-align: center;
+}
+
+#link-md {
+    display: flex;
+    justify-content: space-around;
+}
+
+#link-md a {
+    display: inline-block;
+    vertical-align: middle;
+}
+
+#link-md a {
+    display: inline-block;
+    vertical-align: middle;
+}
+
+#link-md a:nth-child(1) {
+    margin-right: 10px;
+    @media only screen and (max-width: 1000px) { 
+        margin-right: 0px;
+    }
+}
+
+#link-md a:nth-child(2) {
+    margin-left: 10px;
+    @media only screen and (max-width: 1000px) { 
+        margin-left: 0px;
+    }
+}

init.py

@@ -57,7 +57,7 @@ def _initialize_paper_info(source_ds):
             arxivid2data[paper['arxiv_id']] = {"idx": count, "paper": paper}
             date_dict[year][month][day].append(paper)
 
-    titles = title2qna.keys()
+    titles = [f"[{v['arxiv_id']}] {k}" for k, v in title2qna.items()]
 
     return titles, date_dict, arxivid2data
 

ui.py

@@ -13,6 +13,8 @@ from gen.gemini_chat import gen_text as gemini_gen_text
 from gen.gemini_chat import init as gemini_init
 from constants.context import DEFAULT_GLOBAL_CTX
 
+from paper.download import get_papers_from_arxiv_ids
+
 from init import (
     requested_arxiv_ids_df,
     date_dict,
@@ -60,7 +62,10 @@ def get_paper_by_day(year, month, day):
     ))
     return gr.Dropdown(choices=papers, value=papers[0])
 
+# 2307.02040
 def set_papers(year, month, day, title):
+    title = title.split("]")[1].strip()
+
     papers = []
     for paper in date_dict[year][month][day]:
         papers.append(paper["title"])
@@ -82,15 +87,11 @@ def set_paper(year, month, day, paper_title):
             selected_paper = paper
             break
 
-    print(type(selected_paper['arxiv_id']))
-
     return (
         selected_paper['arxiv_id'],
         gr.Markdown(f"# {selected_paper['title']}"), 
         gr.Markdown(
-            "[![arXiv](https://img.shields.io/badge/arXiv-%s-b31b1b.svg?style=for-the-badge)](https://arxiv.org/abs/%s)" % (selected_paper['arxiv_id'], selected_paper['arxiv_id'])
-        ),
-        gr.Markdown(
+            "[![arXiv](https://img.shields.io/badge/arXiv-%s-b31b1b.svg?style=for-the-badge)](https://arxiv.org/abs/%s)" % (selected_paper['arxiv_id'], selected_paper['arxiv_id']) + " "
             "[![Paper page](https://huggingface.co/datasets/huggingface/badges/resolve/main/paper-page-lg.svg)](https://huggingface.co/papers/%s)" % selected_paper['arxiv_id']
         ),
         gr.Markdown(selected_paper["summary"]),
@@ -127,6 +128,8 @@ def set_paper(year, month, day, paper_title):
     )
 
 def set_date(title):
+    title = title.split("]")[1].strip()
+
     for _, (year, months) in enumerate(date_dict.items()):
         for _, (month, days) in enumerate(months.items()):
             for _, (day, papers) in enumerate(days.items()):
@@ -195,8 +198,10 @@ def add_arxiv_ids_to_queue(queue, arxiv_ids_str):
         valid_arxiv_ids = _filter_duplicate_arxiv_ids(valid_arxiv_ids)
 
         if len(valid_arxiv_ids) > 0:
-            valid_arxiv_ids = [[arxiv_id] for arxiv_id in valid_arxiv_ids]
-            gr.Warning(f"Processing on [{valid_arxiv_ids}]. Other requested arXiv IDs not found on this list should be already processed or being processed...")
+            papers = get_papers_from_arxiv_ids(valid_arxiv_ids)
+            valid_arxiv_ids = [[f"[{paper['paper']['id']}] {paper['title']}"] for paper in papers]
+
+            gr.Warning(f"Processing [{valid_arxiv_ids}]. Other requested arXiv IDs not found on this list should be already processed or being processed...")
             valid_arxiv_ids = pd.DataFrame({'Requested arXiv IDs': valid_arxiv_ids})
             queue = pd.concat([queue, valid_arxiv_ids])
             queue.reset_index(drop=True)
@@ -254,13 +259,12 @@ async def chat_stream(idx, local_data, user_prompt, chat_state, ctx_num_lconv=3)
         async for result in gemini_gen_text(prompt):
             ppm.append_pong(result)
             yield "", ppm.build_uis(), str(ppm), gr.update(interactive=False), gr.update(interactive=False), gr.update(interactive=False)
-    
+
         yield "", ppm.build_uis(), str(ppm), gr.update(interactive=True), gr.update(interactive=True), gr.update(interactive=True)
     except Exception as e:
-        print(e)
         gr.Warning(str(e))
         ppm.replace_last_pong("Gemini refused to answer. This happens becase there were some safety issues in the answer.")
-        yield "", ppm.build_uis(), str(ppm), gr.update(interactive=True), gr.update(interactive=True), gr.update(interactive=True)
+        yield "", ppm.build_uis(), str(ppm), gr.update(interactive=True), gr.update(interactive=True), gr.update(interactive=True)        
 
 def chat_reset(local_data, chat_state):
     ppm = chat_state["ppmanager_type"].from_json(json.dumps(local_data))