feat: Minor improvements in usability

src/doc_search/web.py

@@ -14,35 +14,22 @@ from doc_search.workflow import (
 
 pn.extension(loading_spinner="dots", loading_color="#00aa41")
 
-inp = pn.widgets.TextInput(value="", placeholder="Enter text here...")
-button_conversation = pn.widgets.Button(name="Ask me something!")
-convos_text = []  # store all texts in a list
-convos = []  # store all panel objects in a list
+txt_input = pn.widgets.TextInput(value="", placeholder="Enter text here...", sizing_mode="stretch_width")
+btn_ask = pn.widgets.Button(name="Ask me something!", width=100)
+panel_conversations = []  # store all panel objects in a list
 global_context = {}  # store workflow context
 
 
-def get_conversations(_: Any) -> pn.Column:
-    prompt = inp.value_input
-    logging.info("Getting conversation for prompt: %s for input %s", prompt, inp)
-    if prompt != "":
-        input_question = global_context["input_question"] = prompt
-        convos_text.append(input_question)
-        run_workflow(global_context, inference_workflow_steps())
-        openai_answer = global_context["output"]
-        logging.info("Answer: %s", openai_answer)
-        convos_text.append(openai_answer)
-        convos.append(pn.Row("🙂", pn.pane.Markdown(f"**{prompt}**", width=600)))
-        render_answer(openai_answer)
-    inp.value_input = ""
-    return pn.Column(*convos)
-
+def add_qa_to_panel(question: str, answer: str) -> None:
+    qa_block = f"""
+🙂    **{question}**
 
-def render_answer(openai_answer: str) -> None:
-    convos.append(
+📖    {answer}
+    """
+    panel_conversations.append(
         pn.Row(
-            "📖",
             pn.pane.Markdown(
-                openai_answer,
+                qa_block,
                 width=600,
                 style={
                     "background-color": "#F6F6F6",
@@ -53,6 +40,19 @@ def render_answer(openai_answer: str) -> None:
     )
 
 
+def get_conversations(_: Any) -> pn.Column:
+    prompt = txt_input.value_input
+    logging.info("Getting conversation for prompt: %s for input %s", prompt, txt_input)
+    if prompt != "":
+        input_question = global_context["input_question"] = prompt
+        run_workflow(global_context, inference_workflow_steps())
+        openai_answer = global_context["output"]
+        logging.info("Answer: %s", openai_answer)
+        add_qa_to_panel(input_question, openai_answer)
+    txt_input.value_input = ""
+    return pn.Column(*(reversed(panel_conversations)))
+
+
 def run_inference_workflow(context: dict) -> None:
     run_workflow(context, inference_workflow_steps())
 
@@ -62,24 +62,23 @@ def run_web(context: dict) -> None:
     context["index_path"] = pdf_to_index_path(context["app_dir"], context["input_pdf_path"])
     context["faiss_db"] = pdf_to_faiss_db_path(context["app_dir"], context["input_pdf_path"])
     global_context = context
-    interactive_conversation = pn.bind(get_conversations, button_conversation)
+    interactive_conversation = pn.bind(get_conversations, btn_ask)
 
     pdf_name = pdf_name_from(context["input_pdf_path"])
-    convos.append(
-        pn.Row(
-            "📖",
-            pn.pane.Markdown(f"Ask me something about {pdf_name}", width=600, style={"background-color": "#F6F6F6"}),
-        )
+    panel_conversations.append(
+        pn.pane.Markdown(f"📖 Ask me something about {pdf_name}", width=600, style={"background-color": "#F6F6F6"}),
     )
 
     run_workflow(global_context, inference_workflow_steps())
-
-    convos.append(pn.Row("📖", pn.pane.Markdown("**Here is the book summary**", width=600)))
-    render_answer(global_context["output"])
+    add_qa_to_panel("*Here is the book summary*", global_context["output"])
 
     dashboard = pn.Column(
-        inp,
-        pn.Row(button_conversation),
-        pn.panel(interactive_conversation, loading_indicator=True, height=500),
+        pn.Row(txt_input, btn_ask),
+        pn.panel(
+            interactive_conversation,
+            loading_indicator=True,
+            height=500,
+            style={"border-radius": "5px", "border": "1px black solid"},
+        ),
     )
     panel.serve(dashboard, port=5006, show=True)