basic inference pipeline

README.md

@@ -8,6 +8,7 @@ sdk_version: 3.3.1
 python_version: 3.10.4
 app_file: main_hf_space.py
 pinned: false
+license: mit
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

introduck/inference.py

@@ -0,0 +1,43 @@
+import spacy
+
+
+def extract_contacts_from_email(payload: str) -> list[list[str]]:
+    return [["John Doe", "Acme Corp.", "john@example.com", "ai, nn"]]
+
+
+def generate_acceptance_reply(payload: str) -> str:
+    return "Good!"
+
+
+def generate_rejection_reply(payload: str) -> str:
+    return "Not interested :("
+
+
+def highlight_named_entities(payload: str, labels: list[str] = None) -> dict:
+    if labels is None:
+        labels = ["ORG", "PERSON"]
+
+    nlp: spacy.Language | None = None
+    if not hasattr(highlight_named_entities, "nlp"):
+        nlp = spacy.load(name="en_core_web_sm")
+        highlight_named_entities.nlp = nlp
+        print(f"Loaded {nlp.meta.get('name', 'unknown')} model from {nlp.path}")
+    else:
+        nlp = highlight_named_entities.nlp
+        print(f"Reused {nlp.meta.get('name', 'unknown')} model from {nlp.path}")
+
+    doc = nlp(payload)
+
+    entities: list = []
+    for ent in doc.ents:
+        if ent.label_ not in labels:
+            continue
+
+        entity: dict = {
+            "entity": ent.label_,
+            "start": ent.start_char,
+            "end": ent.end_char}
+
+        entities.append(entity)
+
+    return {"text": payload, "entities": entities}

introduck/routes.py

@@ -3,29 +3,14 @@ import gradio as gr
 from fastapi import FastAPI
 from gradio.routes import App as GradioApp
 
-
-def _analyze_message(sender: str, recipients: str, subject: str, message: str):
-    input_message: str = f"""
-    From: {sender or '<empty>'}
-    To: {recipients or '<empty>'}
-    Subject: {subject or '<empty>'}
-
-    {message or ''}
-    """
-
-    contacts: list = [["John Doe", "Acme Corp.", "john@example.com", "ai, nn"]]
-
-    acceptance_reply: str = "Good!"
-
-    rejection_reply: str = "Not interested :("
-
-    return input_message, contacts, acceptance_reply, rejection_reply
-
-
-def _use_message_template() -> (str, str):
-    subject: str = "Could you make an intro?"
-
-    message: str = """\
+from introduck.inference import extract_contacts_from_email
+from introduck.inference import generate_acceptance_reply
+from introduck.inference import generate_rejection_reply
+from introduck.inference import highlight_named_entities
+from introduck.utils import create_email_message
+
+_INTRO_SUBJECT_EXAMPLE: str = "Could you make an intro?"
+_INTRO_MESSAGE_EXAMPLE: str = """\
 Hey,
 
 hope you are doing well! :-)
@@ -45,71 +30,93 @@ Relevant info:
   + <highlights list>
 
 Best,
-<your name>"""
+<your name>\
+"""
+
+
+def _analyze_message(sender: str, recipients: str, subject: str, body: str):
+    msg_data: dict = {
+        "from": sender,
+        "to": recipients,
+        "subject": subject,
+        "body": body}
+
+    msg: str = create_email_message(data=msg_data)
+
+    contacts: list = extract_contacts_from_email(payload=msg)
+    acceptance_reply: str = generate_acceptance_reply(payload=msg)
+    rejection_reply: str = generate_rejection_reply(payload=msg)
+    text_with_named_entities: dict = highlight_named_entities(payload=msg)
 
-    return subject, message
+    return contacts, acceptance_reply, rejection_reply, text_with_named_entities
+
+
+def _use_message_template() -> (str, str):
+    return _INTRO_SUBJECT_EXAMPLE, _INTRO_MESSAGE_EXAMPLE
 
 
 def create_base_route(title: str = "") -> FastAPI:
-    with gr.Blocks(analytics_enabled=False, title=title) as base_blocks:
-        gr.Markdown(f"""\
+    # TODO: Fix once resolved https://github.com/gradio-app/gradio/issues/1683
+    # TODO: ...and remove elem_id="htxt" from HighlightedText for RF822 output
+    css_workaround: str = "#htxt span {white-space: pre}"
+
+    base_blocks: gr.Blocks = gr.Blocks(
+        analytics_enabled=False,
+        title=title,
+        css=css_workaround)
+
+    with base_blocks:
+        gr.Markdown("""\
             # Introduck
-            ## All introductions are better with an AI-powered 🦆
-            No new tools. No complex setup. No hassle.
+            ## Intro requests. Simplified. 🦆
+            No new tools. No complex setup. No hassle.\
             """)
 
-        with gr.Group():
-            email_sender_input: gr.Textbox = gr.Textbox(
-                label="From:",
-                lines=1,
-                max_lines=1,
-                placeholder="founder@acme.com")
-
-            email_recipients_input: gr.Textbox = gr.Textbox(
-                label="To:",
-                lines=1,
-                max_lines=1,
-                placeholder="vc@example.com")
-
-            email_subject_input: gr.Textbox = gr.Textbox(
-                show_label=False,
-                lines=1,
-                max_lines=1,
-                placeholder="Subject...",
-                interactive=True)
-
-            email_message_input: gr.Textbox = gr.Textbox(
-                show_label=False,
-                lines=5,
-                max_lines=42,
-                placeholder="Type a message...",
-                interactive=True)
-
-            # TODO: Add support for attached files.
-            # email_attachments_input: gr.File = gr.File(
-            #     label="Attachments:",
-            #     file_count="multiple")
-
-        with gr.Row():
-            email_template_button: gr.Button = gr.Button(
-                value="Use template",
-                variant="secondary")
-
-            email_template_button.click(
-                fn=_use_message_template,
-                inputs=[],
-                outputs=[email_subject_input, email_message_input])
-
-            email_submit_button: gr.Button = gr.Button(
-                value="Submit",
-                variant="primary")
+        with gr.Box():
+            with gr.Column():
+                gr.Markdown("## New intro request")
+
+                email_sender_input: gr.Textbox = gr.Textbox(
+                    label="From:",
+                    lines=1,
+                    max_lines=1,
+                    placeholder="founder@acme.com")
+
+                email_recipients_input: gr.Textbox = gr.Textbox(
+                    label="To:",
+                    lines=1,
+                    max_lines=1,
+                    placeholder="vc@example.com")
+
+                email_subject_input: gr.Textbox = gr.Textbox(
+                    show_label=False,
+                    lines=1,
+                    max_lines=1,
+                    placeholder="Subject...",
+                    interactive=True)
 
-        with gr.Tabs(selected="default"):
-            with gr.TabItem(label="Input"):
-                rfc822_output: gr.HighlightedText = gr.HighlightedText(
-                    label="RFC822 message:",
-                    show_legend=False)
+                email_body_input: gr.Textbox = gr.Textbox(
+                    show_label=False,
+                    lines=5,
+                    max_lines=42,
+                    placeholder="Type a message...",
+                    interactive=True)
+
+                # TODO: Add support for attached files.
+                # email_attachments_input: gr.File = gr.File(
+                #     label="Attachments:",
+                #     file_count="multiple")
+
+                with gr.Row():
+                    email_template_button: gr.Button = gr.Button(
+                        value="Use template",
+                        variant="secondary")
 
+                    email_submit_button: gr.Button = gr.Button(
+                        value="Submit",
+                        variant="primary")
+
+        with gr.Tabs(selected="default"):
             with gr.TabItem(label="Contacts", id="default"):
                 contacts_table_headers: list[str] = [
                     "Name",
@@ -122,28 +129,21 @@ def create_base_route(title: str = "") -> FastAPI:
                     max_cols=len(contacts_table_headers),
                     max_rows=16)
 
-            with gr.TabItem(label="Accept request"):
-                acceptance_template_output: gr.Textbox = gr.Textbox(
-                    label="Use this message to accept an intro request:",
-                    interactive=True)
+            with gr.TabItem(label="Replies"):
+                with gr.Column():
+                    acceptance_template_output: gr.Textbox = gr.Textbox(
+                        label="Use this message to accept an intro request:",
+                        interactive=True)
 
-            with gr.TabItem(label="Decline request"):
-                rejection_template_output: gr.Textbox = gr.Textbox(
-                    label="Use this message to decline an intro request:",
-                    interactive=True)
+                    rejection_template_output: gr.Textbox = gr.Textbox(
+                        label="Use this message to decline an intro request:",
+                        interactive=True)
 
-        email_submit_button.click(
-            fn=_analyze_message,
-            inputs=[
-                email_sender_input,
-                email_recipients_input,
-                email_subject_input,
-                email_message_input],
-            outputs=[
-                rfc822_output,
-                contacts_output,
-                acceptance_template_output,
-                rejection_template_output])
+            with gr.TabItem(label="RFC822"):
+                rfc822_output: gr.HighlightedText = gr.HighlightedText(
+                    label="RFC822 message (for debugging purposes):",
+                    show_legend=False,
+                    elem_id="htxt")  # temporary fix, see above for more info
 
         gr.Markdown("## FAQ")
 
@@ -162,6 +162,26 @@ def create_base_route(title: str = "") -> FastAPI:
 
         gr.Markdown("**&copy;** Introduck, 2022")
 
+        # ui end, continue with signals and slots...
+
+        email_template_button.click(
+            fn=_use_message_template,
+            inputs=[],
+            outputs=[email_subject_input, email_body_input])
+
+        email_submit_button.click(
+            fn=_analyze_message,
+            inputs=[
+                email_sender_input,
+                email_recipients_input,
+                email_subject_input,
+                email_body_input],
+            outputs=[
+                contacts_output,
+                acceptance_template_output,
+                rejection_template_output,
+                rfc822_output])
+
     base_route = GradioApp.create_app(blocks=base_blocks)
 
     # TODO: overwrite /favicon.ico

introduck/utils.py

@@ -0,0 +1,16 @@
+from email.message import EmailMessage
+
+
+def create_email_message(data: dict) -> str:
+    msg: EmailMessage = EmailMessage()
+
+    body_text: str = data.get("body", "")
+    # body_html: str = ""
+
+    msg["From"] = data.get("from", "")
+    msg["To"] = data.get("to", "")
+    msg["Subject"] = data.get("subject", "")
+    msg.set_content(body_text)
+    # msg.add_alternative(body_html, subtype="html")
+
+    return msg.as_string()

requirements.txt

@@ -1,3 +1,6 @@
 fastapi[all]
 gradio==3.3.1  # keep in sync with version from README.md metadata
+spacy==3.4.0
 uvicorn[standard]
+
+https://huggingface.co/spacy/en_core_web_sm/resolve/main/en_core_web_sm-any-py3-none-any.whl