This repository contains various bit of half baked projects around the concept of define html directly in python. This is my hobby project and as such it is in no way organised.
The design guide that I have used for the various projects is that of minimal externally installed dependencies. This means that if I need an external library it should at most be one file dependency. This was because I needed it portable on a usb flash stick which would be used on both Linux and Windows. So it would have been difficult to use Virtual Environment.
Each project is contained within it own folder as a module which can be run using python -m <project_name>.
The only folder that is not a project is the tools folder which contains all the custom framework code that is used in the each project. The reason for the name tools is for the unimaginative reason of diffuculty in naming it.
Of course the project is not without its external dependencies.
- (LanguageTool (for the grammar server) in java)[https://languagetool.org/] not required can be disabled
- (NLTK - nlp)[nltk.org] (Not required)
- (Matplotlib)[https://matplotlib.org/] - for specific projects
: Note that the working directory needs to be the root of the git repo.
This project is my custom tooling for writing a creative work in the form of a story. Orginally used flask as its server framwork but then ported to (bottlepy - a single file web-framework)[https://bottlepy.org/]. The story behind is that tradtional editing such as markdown or LibreOffice were not flexible enough to do what I needed which was embed references to varouis story elements so that I could hover over or click on a referenced word and it would pull up information about the reference. This was needed for note taking. This project is by far the most featurefull.
server3/test.py was my attempt at performance analysis of transforming Element() to str
- NLTK - sentence tokenising (Not required)
- Only when used along side LanguageTool and for serialisation in
server3/serialize.py (run as module)
- Only when used along side LanguageTool and for serialisation in
grammar checking can be toggled in server3/config.py
This project is a probably a secondary goal of the entire frameworks minimal large dependencies. This project
archive the entire tools package along with its own source code and serves a webpage that runs
(pyscript)[https://pyscript.net/] that loads the archive and using the tools.htmls module generates the entire
web page in the browser in python. Of this was implemented to eases my development process but could for production
use python wheels instead. It is my hope that this would bring python into the webbrowser.
Another feature which could be achieved is constructing a virtual dom out of python Element() with say a
reference to html objects in the instance. This is the purpose of Element_Constructor which would allow
the base class of all Elements to be tailored for the browser or perhaps a desktop gui framework like Qt or gtk.
This was my attempt at remote function calling but with added advantage of interfaces that are defined in the source
rather than at runtime like xmlrpc in the standard library so that code completion workds. There are three
major problems with problems with the implementation first is security, the protocol shunts around pickled objects.
Second is method chaining it not possible to rpc.method1().some_property as this needs to be send two requests to
the server but python can't handle it and errors. Finnaly, to the best of my knowledege the server code is
able to handle request - process - response in a single request but it takes multiple request to feed the response
back up the chain.
Quite self explanatory. I think I followed a tutorial somewhere not sure where though.
- (Peewee ORM - single file)[http://docs.peewee-orm.com/en/latest/] -
tools/dbms/peewee.py
This project has been in the pipeline for a long time and been rewritten serveral times. It is a Recipe book editor, viewer and framework. Started out as a simple gui written in tkinter to generate ReStructuredText for sphinx. Then a PyQt application with xml recipe storage. Now it is a fully pythonized storage system of pickling objects. It is an attempt to use what I call the Provider Pattern where a Provider object is created and DataStores are registered on it. Thus allowing for a single collated view of multiple sources.
The reason for this is say the user enters some ingredients before a cookbook is loaded then those ingredients are stored in the application user-data folder or a more concretely the application can provide a datastore of commonly used ingredients that are then collated along with cookbook specific ingredients so that the user would not need to re-enter all the ingredient data each time a new a cookbook is created.