O(1) Memory Streaming

One of Probo UI’s most powerful features is its ability to stream HTML directly to the browser with almost zero memory overhead.

Traditional template engines (Jinja2, Django Templates, etc.) build the entire HTML string in memory before sending it. This causes high RAM usage and slow Time-to-First-Byte (TTFB) when rendering large datasets.

Probo UI solves this with true O(1) memory streaming using Python generators.

---

Why Streaming Matters

Rendering a table with 50,000+ rows using .render() can: - Spike your server’s RAM - Delay the first byte sent to the user - Make your application feel slow

Probo’s .stream() method fixes this by yielding HTML in small chunks as it’s generated.

---

Standard vs Streaming Approach

❌ Using .render() (High Memory)

from probo import ul, li

# This loads everything into memory at once
users = get_all_users()                    # Big list in RAM
ui = ul(*[li(user.name) for user in users]) # Another huge list + HTML string

html = ui.render()                         # Server freezes

✅ Using .stream() (O(1) Memory)

from probo import ul, li

# This loads everything into memory at once
users = get_all_users()                    # Big list in RAM
ui = ul(*[li(user.name,return_list=True,stream=True) for user in users],return_list=True,stream=True) # Another huge list + HTML string

# Stream chunks to the client
for chunk in ui:
    send_to_client(chunk)

Using .stream() in wild

from probo import TABLE,tr,td
from .models import User

def gen_user_table(*args,**kwargs):
    users = User.objects.all().iterator()

    lazy_rows = (tr(td(u.id), td(u.name)) for u in users)

    ui = TABLE(lazy_rows)
    return ui

• Memory usage stays low because rows are processed and sent one batch at a time.

Integration with Web Frameworks

• Django Example (Using StreamingHttpResponse)

from django.http import StreamingHttpResponse
from .components.users import gen_user_table

def streamed_users_view(request):
    ui = gen_user_table()
    ui.style_manager.table_layout= "fixed"
    ui.style_manager.width= "100%"

    return StreamingHttpResponse(
        ui.stream(batch=50),
        content_type="text/html"
    )