syntaxmatrix 1.3

SyntaxMUI: A customizable UI framework for Python AI Assistant Projects.

syntaxmatrix UI Framework v1.2.0

SyntaxMUI: A customizable UI framework for Python AI assistant projects.

---

Overview

syntaxmatrix (lowercase) is a lightweight, Pythonic UI library that lets you build interactive chat-style front-ends for AI apps without diving into a full web framework. It provides:

• Rapid widget registration (text inputs, buttons, file uploaders)
• Dynamic theme toggling and multiple UI modes
• Built-in PDF ingestion & chunking
• Stylized feedback (success, error, warning, info)
• Admin panel, session management, and more

Ideal for Retrieval-Augmented Generation (RAG) demos, data explorers, or any AI assistant interface.

---

Quick Links

• Documentation: https://github.com/bobganti/SyntaxMatrix
• PyPI: pip install syntaxmatrix==1.2.1
• Source: https://github.com/bobganti/SyntaxMatrix

---

Features

• Rapid UI Creation: One-line calls to register text_input, button, file_uploader, etc.
• Built-in Chat Loop: Core /process_chat endpoint automatically wires three keys:
  • user_query (text input),
  • submit_query (send button),
  • user_pdfs (PDF uploader + chunking).
• Custom Widgets: Register any keys you like—just provide matching handler logic.
• PDF Ingestion: load_pdf_chunks(directory) scans a folder, splits PDFs into chunks, caches in SQLite, and returns a {filename: [chunks,…]} map.
• Session File Uploads: get_user_chunks(), add_user_chunks(), clear_user_chunks() for per-session user files.
• Dynamic Themes: enable_theme_toggle(), set_theme(), plus list_themes().
• Multiple UI Modes: set_ui_mode() supports default, bubble, card, smx (new!) and list_ui_modes().
• Rich Output: markdown(), latex(), plt_plot(), plotly_plot(), plus error(), warning(), success(), info().
• Branding Helpers: set_user_icon(), set_bot_icon(), set_site_icon(), set_site_logo(), set_site_title(), set_project_title().
• Session Management: Named chat sessions with rename/delete; session IDs via get_session_id().

---

Installation

pip install syntaxmatrix==1.2.0

Dependencies

• Python >= 3.7
• Flask >= 2.0.0
• requests >= 2.0.0
• markdown >= 3.3.0
• matplotlib >= 3.5.0
• plotly >= 5.0.0
• openai >= 0.27.0
• PyPDF2 >= 1.26.0
• numpy >= 1.21.0

Optional extras:

# Transformer models & torch
pip install syntaxmatrix[advanced_nlp]

# pandas for data APIs
pip install syntaxmatrix[data]

# Testing
pip install syntaxmatrix[testing]

---

Quick-Start Snippet

This minimal example uses the built-in chat flow (no custom keys required):

import syntaxmatrix as smx

# Handler for text-submission
def create_conversation():
    q = smx.get_text_input_value("user_query").strip()
    if not q:
        smx.warning("Please enter a question.")
        return
    # …your AI call or logic here…
    smx.success(f"You asked: {q}")

# UI setup
smx.set_ui_mode("smx")               # new "smx" display style
smx.enable_theme_toggle()            # add light/dark toggle

# Activate built-in widgets (must use these keys)
smx.text_input(
    "user_query",                  # text box
    "Ask me anything…",            # label
    placeholder="Type your question here…"
)
smx.button(
    "submit_query",                # send button key
    "Send",                        # label
    callback=create_conversation    # invoked on send
)
smx.file_uploader(
    "user_pdfs",                   # PDF uploader key
    "Upload PDFs",                 # label
    accept_multiple_files=True     # chunking happens automatically
)

if __name__ == "__main__":
    smx.run()

Note: If you register different keys, you must also adapt your handlers and/or the request-processing logic to match those names.

---

Registering Custom Widgets

You’re not limited to the three defaults. Example:

# Custom keys:
smx.text_input("query_box", "Your query…")
smx.button("ask_btn", "Ask", callback=my_ask_handler)
smx.file_uploader(
    "pdf_docs",
    "Attach Documents", 
    accept_multiple_files=True,
    callback=my_upload_handler
)

Then your my_ask_handler() must read from smx.get_text_input_value("query_box"), and you’ll need to implement or wrap the chat loop yourself if you deviate from the built-ins.

---

API Reference

App Lifecycle & Branding

Function	Description
run()	Launches the Flask server and opens a browser window
set_site_title(title: str)	Sets the navbar site title
set_project_title(title: str)	Sets the main project heading
set_user_icon(icon: str)	Emoji/text for user messages
set_bot_icon(icon: str)	Emoji/text for bot messages
set_site_icon(icon: str)	Small icon in the browser tab
set_site_logo(logo: str)	Text/logo in the navbar

Theming & Modes

Function	Description
enable_theme_toggle()	Show a light/dark toggle link in the navbar
disable_theme_toggle()	Hide the theme toggle
set_theme(name: str, theme_dict?: dict)	Switch or define a custom theme
list_themes() -> List[str]	Get all available theme names
set_ui_mode(mode: str)	Choose layout: default, bubble, card, smx
list_ui_modes() -> Tuple[str,…]	Available UI modes

Built-in Widgets & Flow

Function	Description
text_input(key, label, placeholder="")	Register a text box. For built-in chat use key=user_query.
button(key, label, callback=None)	Register a button. For built-in chat use key=submit_query.
file_uploader(key, label, accept_multiple_files=False, callback=None)	Register a file upload. For PDF chunking use key=user_pdfs.
get_text_input_value(key)	Read current text in a box
clear_text_input_value(key)	Clear the text box
get_file_upload_value(key)	Access raw file objects uploaded

PDF & File-Chunk APIs

Function	Description
load_pdf_chunks(directory: str = "uploads/sys")	Ingest all system PDFs → split into chunks → cache → return {file:[chunks]}
get_session_id() -> str	Current chat session UUID
add_user_chunks(sess_id: str, chunks: List[str])	Store user-uploaded text chunks
get_user_chunks(sess_id: str) -> List[str]	Retrieve stored user chunks
clear_user_chunks(sess_id: str)	Remove all user chunks for session

Rich & Stylized Output

Function	Description
write(content: str)	Append raw HTML/text to system-output buffer
markdown(md_text: str)	Render Markdown via Python-Markdown
latex(math_text: str)	Render LaTeX math via MathJax
error(msg: str)	Output red-styled error message
warning(msg: str)	Output orange-styled warning
success(msg: str)	Output green-styled success
info(msg: str)	Output blue-styled info
plt_plot(fig: matplotlib.figure.Figure)	Embed a Matplotlib figure
plotly_plot(fig: plotly.Figure)	Embed a Plotly figure

---

License

MIT © Bob Nti