Python Installation
Title is a bit misleading as these are not general notes, but notes on how I install and how you might get some of my applications ( see the categoy below on Python Projects 0)
Contents
Basic Install[edit]
I use the Anaconda installation because it gives a wide range of Python Packages, A Package and Virtual Environment Manager, a very good Python IDE ( Spyder ), IPython, Juypter Notebooks... all in one install. Read below all the way through.
I have links at the bottom of the page.
- Download
- Install
- Run Anaconda from link on desktop ( if one is missing in window, type Navigator, right click, say open folder, say send shortcut to desktop )
- But install often goes to odd location like the hidden folder ProgramData. This can cause ownership and permission issues. So I pick my own directory like c:/apps/Anaconda ... i own this, it is shorter, and there are no spaces in the full path name. But do what you like best.
Maintenance and Additions[edit]
[edit]
The second page of Navigator called Environments is a nice way to create, modify and delete Python virtual environments. On the left hand side you can choose your environment, and on the right hand side you can manage packages for the given environment. If that does not entirely work for you, you can just pick your environment on the left and have it drop you to a command line ( terminal ) for that environment. At that command line you can run python, conda, pip, or whatever you want.
Using Conda[edit]
I start with Navigator to get to a terminal in the right Python environment then use conda. Conda is pretty much like pip ( conda install something_wonderful ) but different. It is supposedly not full compatible with pip so I only use pip when conda does not work. Sometimes the command are a bit more complicated as: "conda install -c conda-forge spacy" which I use to install spacy.
Using Pip[edit]
When conda fails use pip.
Using Setup.py[edit]
When pip fails consider setup.py, usually will have a readme file in the package for your guidance... after this there is still google or the author of the software you are using this package for.
Some Packages I use -- Not in the Anaconda Distribution[edit]
I always used conda first and if that does not work move on to pip if that does not work perhaps setup.py
Pyperclip[edit]
- Why: For Python to access the system clipboard. Get Copy and Paste working in your applications
- Install: pip install pyperclip
- May be used in Clipboard, EasyDB, TwitterDB () SmartTerminal
Spacy[edit]
- Why: For Python to process language into its base words or lemmas
- Install: with conda, but from command line. Directions at: Spacy :: Anaconda Cloud Command I used: "conda install -c conda-forge spacy" After the install my code still had issues which I fixed with: "python -m spacy download en_core_web_sm" for more info see: spacy Can't find model 'en_core_web_sm' on windows 10 and Python 3.5.3 :: Anaconda custom (64-bit) - Stack Overflow
- May be used in TwitterDB
- Links: Lots of info out there on spacy, google it.
HTML[edit]
- Why: For Python to process language into its base words or lemmas
- This is an old package that I came across and liked, there may be better ones, but this did what I wanted for a couple of my applications. It cannot currently be installed with conda or pip. All the code, which I downloaded and decompressed is in each application where I use it ..../resources/HTML.py-0.04 The install is documented in the readme file, I used "python setup.py install" from the source code (HTML) directory.
- May be used in EasyDB, TwitterDB
- Links
Working With Someone Else's Source[edit]
So follow their directions. If it is my code this is pretty minimal. I have not started using requirements.txt but will if I get many requests. In any case here is pretty much the last resort method:
- Run the code.
- Stop when an import fails. Try to install with conda.... pip.... An unfortunate fact is that the package name for conda/pip is not always exactly like the import name ( but usually is ), so a bit of google may help.
- Usually after an install you can just restart the program, but I have sometimes found that restarting an IDE is necessary.
- Repeat with the next import that fails.
Anaconda Links[edit]
- FAQ — Conda documentation
- Anaconda Overview | Continuum
- Presentations & Blog Posts — Conda documentation
- Python « Steve Byrnes's Homepage
- Installation — Spyder 2.3 documentation
- Installation of Python, Spyder, Numpy, Sympy, Scipy, Pytest, Matplotlib via Anaconda (2013) — Computational Modelling Blog
- spyder
- Anaconda FAQ | Continuum Analytics: Documentation
- Anaconda install | Continuum Analytics: Documentation
- Download Anaconda Now! | Continuum
- Installing Jupyter Notebook — Jupyter Documentation 4.1.1 alpha documentation