grepsr-cli 0.9.39

A Cli tool for Grepsr Developers

A Cli tool for Grepsr Developers

Installation

$ pip install grepsr-cli

Usage

passing parameters to amazon_com service.

gcli crawler test -s amazon_com -p '{"urls":["https://amazon.com/VVUH4HJ","https://amazon.com/FV4434"]}'

if JSON is complex, use file instead

# contents of /tmp/amazon_params.json
{"urls": ["https://amazon.com/VV%20UH4HJ"], "strip": ["'", "\"", "\\"]}

gcli crawler test -s amazon_com --params-file '/tmp/amazon_params.json'

Hacks Used.

If the json parameter has a space, it might break parameter parsing. If the json parameter has a dash - and any character after it has a space, it will break parameter parsing. Cause: no double quoting around $@ in run_service.php:5:49 here This is fixed hackily by replacing string with its unicode \u0020 sequence. This works beacause $@ does not split on \u0020.

inject custom command.

Say, for example you wanted to a inject a php function so that it could be called from inside you service_code when testing locally. Note: All these files should only be created inside ~/.grepsr/tmp. Creating it outside will not work.

1. Create a file called inject.php inside ~/.grepsr/tmp/
2. Implement your function inside ~/.grepsr/tmp/inject.php

function addRowLocal($arr) {
    ...
    ...
}

3. Create a file called inject.sh inside ~/.grepsr/tmp/
4. inside inject.sh add:

alias php='php -d auto_prepend_file=/tmp/inject.php'

Note: the file location is /tmp/inject.php instead of ~/.grepsr/tmp/inject.php. This is because, the local path ~/.grepsr/tmp gets mapped to /tmp in the docker container. And inject.sh runs inside docker, instead of the local filesystem. 5. Add an entry in ~/.grepst/config.yml like so:

    php:
        ...
        sdk_image: ...
        pre_entry_run_file: inject.sh      # relative and limited to the tmp/ dir

6. Now you can use addRowLocal() in your any of your files.

public function main($params) {
    ...
    $arr = $this->dataSet->getEmptyRow();
    addRowLocal($arr); // won't throw error
    ...
}

Development

Be sure to uninstall gcli first, with pip uninstall grepsr-cli

git clone [e-mail]:zznixt07/gcli.git grepsrcli
cd grepsrcli
pip install -e .

Features Added

• drop stash after pushed successully. Before this, all stashes were always kept.
• run a custom shell file before running your crawler. This allows possiblity like always injecting a php function in all your crawlers.
• auto add Dependencies: ... that your crawler class extends (dependecies that are not extended by crawler classes but used elsewhere is upcoming)

TODO:

• Experiment with git rebase on deploy fail. git rebase origin/master --autostash && git push
• Handle Prioritization of same plugin name across multiple repo more deterministically. (maybe prioritize cwd path?)
• node only run crawler if npm install is successfull. (add && between npm install and npm start)
• run tsc before deploying vortex-ts-registry packages
• add option to force update dependecies to latest version for all/specific vortex-ts-registry dependencies
• handle ctrl+c during node package install on docker. (currently it continues running in BG)
• add new baseclass typescript package and do not include SOP, (do not normalize - to _) change npm start to tsc and test runs. Also generate .d.ts file in tsconfig.json