Skip to main content

Displays Polus Render dashboard in Jupyter Notebooks.

Project description

Polus Render

Render application is loaded in an iframe. The package allows pointing the iframe at:

  • Render deployed to a server
  • A Python server running on localhost and serving a production build of render, which has been bundled with this package

The are three ways to load the data:

  1. Specifying a URL to a server
  2. Specifying a local path will start a Python server on localhost. The URL to the dataset on localhost will be passed to the application in the iframe
  3. Dragging-and-dropping the dataset does not use a server, it calls an API from the front end (It should the this under the hood https://developer.mozilla.org/en-US/docs/Web/API/File_API)

Has the ability to work both local and remote versions of Jupyter Notebook.

image

Requirements

Installation

pip install polus-render

Dev Installation

git clone https://github.com/jcaxle/polus-render.git
cd polus-render
[optional] python -m venv venv
[optional] "venv/Scripts/Activate"
pip install -r requirements.txt

Optional steps refer to setting up venv which is recommended.

Project File Structure

polus-render
| Build Instructions.md           // Instructions on how to update Pypi project
| MANIFEST.in                     // Packaging entries
| pyproject.toml                  // Pypi config 
| README                          
| requirements.txt
└───src
    | polus.py                    // Main file, contains render function used by user
    | polus-render-wrapper.py     // Unused file used for a scrapped project. Can be used as a reference for input sanitization
    | zarr_file_server.py         // Contains server only used for serving local build of Polus Render
    ├───apps           
    │   ├───render-ui              // Build files of Polus Render
    │   └───updog-render           // Server used for serving files.

Build Instructions

Submodules

Render: Local build vs online

polus-render is bundled with a build of Polus Render which supports additional functionality compared to the web version. Table is accurate as of 10/4/2023.

Version Zarr from URL/Path TIF from URL/Path Micro-JSON Support Zarr/TIF Drag & Drop Micro-JSON Drag & Drop
Local :heavy_check_mark: :heavy_check_mark: :heavy_check_mark: :heavy_check_mark: :heavy_check_mark:
Online :heavy_check_mark:

Drag & Drop Demo

ezgif-4-7162ca42b5

Local Jupyter Notebooks Demo

TODO

Remote Jupyter Notebooks Demo

TODO

Sample usage

from polus import render, nb_render

# pathlib and urllib are built-ins
from urllib.parse import urlparse
from pathlib import Path

# Embeds an IFrame of a local build of Polus Render into Jupyter Notebooks
render()

# Embeds an IFrame of Polus Render into Jupyter Notebooks
render(use_local_render=False)

# Embeds an IFrame of a local build of Polus Render with an image file hosted at "https://viv-demo.storage.googleapis.com/LuCa-7color_Scan1/"
render(image_location=urlparse("https://viv-demo.storage.googleapis.com/LuCa-7color_Scan1/"))

# Embeds an IFrame of a local build of Polus Render with an image hosted locally at "C:\Users\JeffChen\OneDrive - Axle Informatics\Documents\zarr files\pyramid.zarr"
render(image_location=Path(r"C:\Users\JeffChen\OneDrive - Axle Informatics\Documents\zarr files\pyramid.zarr"))

# Embeds an IFrame of a local build of Polus Render with an image and overlay file that is hosted locally
render(image_location=Path(r"C:\Users\JeffChen\OneDrive - Axle Informatics\Documents\zarr files\pyramid.zarr"), \
microjson_overlay_location=Path(r"C:\Users\JeffChen\OneDrive - Axle Informatics\Documents\overlay files\x00_y01_c1_segmentations.json"))

# Embeds an IFrame of a local build of Polus Render with an image and overlay file that is hosted online
render(image_location=urlparse("https://files.scb-ncats.io/pyramids/segmentations/x00_y01_c1.ome.tif"), \
microjson_overlay_location=urlparse("https://files.scb-ncats.io/pyramids/segmentations/x00_y03_c1_segmentations.json"))

# Embeds an IFrame with a height of 1080 of a local build of Polus Render.
render(height=1080)

# Embeds an IFrame into remote jupyter notebooks. Use this function with argument nbhub_url to specify your notebooks url which must have lab in its url
nb_render(nbhub_url=urlparse("https://jh.scb-ncats.io/user/jeff.chen@axleinfo.com/user-namespaces/lab?"), image_location=Path("work/pyramid.zarr"))

Functions

def render(image_location:Union[ParseResult, PurePath] = "", microjson_overlay_location:Union[ParseResult, PurePath] = "", width:int=960, height:int=500, image_port:int=0, \
           microjson_overlay_port:int=0, use_local_render:bool=True, render_url:str = "https://render.ci.ncats.io/")->str:
    """
    Displays Polus Render with args to specify display dimensions, port to serve,
    image files to use, and overlay to use.
    
    Param:
        image_location(ParseResult|Purepath): Acquired from urllib.parse.ParseResult or Path, renders url in render.
                            If not specified, renders default render url.
        microjson_overlay_location(ParseResult|Purepath): Acquired from urllib.parse.ParseResult or Path, renders url in render.
                            If not specified, renders default render url
        width (int): width of render to be displayed, default is 960
        height (int): height of render to be displayed, default is 500
        image_port (int): Port to run local zarr server on if used (default is 0 which is the 1st available port).
        microjson_overlay_port (int): Port to run local json server on if used (default is 0 which is the 1st available port).
        use_local_render (bool): True to run local build of render with 1st available port, False to use render_url (default is True)
        render_url (str): URL which refers to Polus Render. Used when run_local_render is False. (default is https://render.ci.ncats.io/)
    Pre: zarr_port and json_port selected (if used) is not in use IF path given is Purepath
    Returns: Render URL
    """

def nb_render(nbhub_url:ParseResult,image_location:Union[ParseResult, PurePath] = "", microjson_overlay_location:Union[ParseResult, PurePath] = "", width:int=960, height:int=500, \
            use_local_render:bool=True, render_url:str = "https://render.ci.ncats.io/")->str:
    """
    Variant of render() used for remote jupyter notebooks. Read render() for usage information

    Param:
        nbhub_url: URL used used for jupyterhub. Contains '/lab/' in its uri
        image_location(ParseResult|Purepath): Acquired from urllib.parse.ParseResult or Path, renders url in render.
                            If not specified, renders default render url.
        microjson_overlay_location(ParseResult|Purepath): Acquired from urllib.parse.ParseResult or Path, renders url in render.
                            If not specified, renders default render url
        width (int): width of render to be displayed, default is 960
        height (int): height of render to be displayed, default is 500
        run_local_render (bool): True to run local build of render with 1st available port, False to use render_url (default is True)
        render_url (str): URL which refers to Polus Render. Used when run_local_render is False. (default is https://render.ci.ncats.io/)
    Returns: Render URL
    """

Implementation Details

  • render() receives sanatized input. Check polus-render-wrapper.py or sample usage for examples on sanitizing input.
  • render() builds up URL scheme fragments for render url, image url, and microjson url.
  • If the image url and microjson url are file paths, serve the files on file servers pointing to either user specified port or a free port.
  • If local render is used, build a server for it as well.
  • At the end, combine render url fragments into a single url, insert it into an IFrame, and display it.
  • Complete url string is returned not printed.
  • For nb_render(), no servers are launched. Files are served from endpoints generated from the remote Jupyter Notebook's URL. The local build of render is served from the polus-server-ext instead of the bundled build files. Essentially, nb_render() runs serverless with the exception of the server that serves the remote Jupyter Notebooks itself.

Misc Implementation Details

  • Two type of servers are used.
  1. Python HTTPServer with CORS and OPTIONS functionality to serve RenderUI
  2. Modified UpDog Flask server to serve local files to RenderUI

Acknowledgements

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

polus-render-0.0.4.0.1.4.tar.gz (4.9 MB view hashes)

Uploaded Source

Built Distribution

polus_render-0.0.4.0.1.4-py3-none-any.whl (5.0 MB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page