Stock-Pattern

A Python CLI scanner to detect and plot common chart patterns

Supports Python >= 3.8 (As of v1.1.1-alpha)

If you ❤️ my work so far, please 🌟 this repo.

These were a series of pattern scanners I wrote for experimentation and fun in 2022. My inspiration was this Medium article Algorithmically Detecting (and Trading) Technical Chart Patterns with Python

See Wiki to understand how the code works: Pattern Algorithms

Important Info

All patterns are detected, prior to breakout. at the last leg of the pattern. Add the stock to your watchlist and track them further.

This program does not provide any buy or sell signals. It only detects the pattern. It is upto you to decide, if a pattern is valid or tradeable.

Installation

Stock-Pattern is being constantly updated with new features and bug fixes. Run git pull to get the latest updates.

1. Clone or download the repo: git clone https://github.com/BennyThadikaran/stock-pattern.git
2. Install dependencies: pip install -r requirements.txt
3. cd into /src folder, run py init.py. It will generate a user.json file
4. Open user.json and edit the DATA_PATH with the folder path to your OHLC data. Files must be in CSV format in any timeframe.
   • If using EOD2 point it to src/eod2_data/daily
   • Optionally, add a SYM_LIST with a file path (CSV or TXT file) containing a list of symbols to scan. (One on each line). It will serve as a default watchlist to scan. See Usage
   • Windows users: add an extra backslash for file paths to avoid JSON decode errors. \\Documents\\python\\stock-pattern
5. See Configuration for more options

{
  "DATA_PATH": "~/Documents/python/eod2/src/eod2_data/daily",
  "POST_SCAN_PLOT": true,
  "SYM_LIST": "./nifty_500.csv"
}

Usage

Display Help: py init.py -h

Show Version: py init.py -v

py init.py (-f filepath | --sym SYM [SYM ...] | -v) options

Scan stocks using a watchlist file. Pass a file containing stocks (one on each line).

# starts an interactive prompt
py init.py -f nifty_500.csv

Scan results are stored in a json file. To plot the results, see Chart Plot options

Scan stocks from a list of symbols. Pass a list of symbols space separated.

py init.py --sym tcs astral datapattns

NOTE: If SYM_LIST is specified in user.json, -f or --sym is not required unless you wish to override the default watchlist.

Identify chart patterns on a specific date.:

# iso date format YYYY-MM-DD
py init.py -f nifty_500.csv -d 2023-01-01

Save results as images to folder

# default folder is src/images
py init.py -f nifty_500.csv --save

# add a custom folder
py init.py -f nifty_500.csv --save ~/Desktop/pattern/

To skip the interactive prompt, specify a pattern string:

# Bullish VCP
py init.py -f nifty_500.csv -p vcpu

Pattern string can be one of:

Pattern String	Description
all	All patterns
bull	All Bullish patterns
bear	All Bearish patterns
vcpu	VCP Up (Bullish)
vcpd	VCP Down (Bearish)
dbot	Double Bottom
dtop	Double Top
hnsu	Head & Shoulder Up (Bullish)
hnsd	Head & Shoulder Down (Bearish)
trng	Triangles (Symmetric, Ascending, Descending)

There are other CLI options available for those wanting to tinker.

Chart Plot options

When patterns are detected the results are saved as json files. Filename is same as pattern string, for Ex. vcpu.json

To plot results of scan, use --plot.

py init.py --plot vcpu.json

To jump to a particular count in the result, use --idx

py init.py --plot vcpu.json --idx 15

When scanning using a watchlist file, all patterns are tracked in json files stored in src/state folder. Files format is <watchlist_name>_<pattern_name>.json for ex. nifty_500_vcpu.json.

Only newly detected patterns are notified and plotted on charts.

To see all currently active patterns in market, use the --plot passing the json file from state folder

py init.py --plot state/nifty_500_vcpu.json

Chart keyboard navigation

Key	Description
n	Next Chart
p	Previous Chart
NUM_KEY + j	Type the number followed by j to jump to index. Press ESC to clear

Configuration

user.json is used for configuration and is auto-generated on first run.

Both absolute and relative file path are supported. Tilde or ~ for home directory symbol is also supported.

Parameter	Description
DATA_PATH	Required. Folder path for OHLC csv data.
SYM_LIST	Optional. File path containing list of symbols, one per line. Override with -f or --file
SAVE_FOLDER	Optional. Folder path to save charts as images. Override with --save
POST_SCAN_PLOT	Default False. If True, plots the results on the chart after a scan.

Screenshots

HUDCO - Double Bottom 16th June 2022

GRAPHITE - Bear VCP 18th Oct 2022

MGL - Reverse Head & Shoulder 16th June 2022

GSFC - Triangle 13th March 2023

RECLTD - Double Tops 13th Oct 2021

TODO

• Make package available via pip [Delayed].

Disclaimer

This program identifies common chart patterns and does it reasonably well (IMHO).

Stock-Pattern is not aimed at replacing a trader's trained eye. It is merely an assistance and learning tool.

CLI options available, for you to backtest the patterns. If you intend to trade these patterns, study them well and trade at your own risk.