oipulse-alerting

Alerts when trending oi data reaches to specified threshold

Sound used in this code is taken from the show Silicon Valley Youtube: https://www.youtube.com/watch?v=uS1KcjkWdoU

Install all python packages in requirements.txt

pip install -r requirements.txt

Environment file setup for trending OI alerting

Define environment variables in .env file

Example: 

AUTH_TOKEN="<your_token_here>" # Get Auth token from oi pulse website after login
EXPIRY="211202" # Expiry in format YYMMDD
TIME_FRAME=5  # Time frame for calculation 3, 5, 10, 15, 30 or 60
DIFF_THRESHOLD=3000000  # Trending OI Difference threshold for alert

Load environment variables in your shell

export `cat .env | xargs`

Running the script

./run_scripts/trending_oi_alerts.sh

Notes

trending_oi.py script generates the alert only when threshold oi difference is reached
Calculations can be little different from trending oi user interface due to different timestamp used
Can use any timeframe for alerting and data will be fetched at particular given intervals for oi difference calculation.

OSPL Signal Alerting

This is a temporary solution implemented using tamper monkey extension and pytesseract python package
It detects the signal based upon image quality and clarity but not 100% perfect though
Limitations include unable to read if 2 signals overlap on each other but still alert will come based upon new entry number detected
Also, signals are best detected when background is black on chart

How to use OSPL signal script?

First install the tamper monkey script on firefox,
Open advance charts and do settings as per image show below
Click Start/Stop button on top left corner to start/Stop capturing the canvas images
Run the python script ./run_scripts/ospl_signal_alerts.sh
Change the download file location in firefox to location of this project stored in our machine
Change the setting to directly allow downloading file without prompt in firefox

Advance Chart settings and Start/Stop button:

OSPL Signal Alerting working explanation:

It collects the canvas images at 4 second interval using tamper monkey script on firefox
Canvas images are stored directly into project folder filename ospl_image.txt
Python script uses this canvas image to OCR signal data using pytesseract package
Current day fetched signals are saved in filename signals_<current_date>.txt
Alerts are generated with 2 seconds delay and only if tesseract detects any new signal.

**Not a perfect solution but a good workaround :)

Note: code tested only on MacOS, feel free to contribute :)

For Education purposes only

Harpreet4u/oipulse-alerting