We Keep Coding

Open Alteryx .yxdb file in Python?

Is there a way to import .yxdb (Alteryx database files) into Pandas/Python, without using Alteryx as a go-between?

The short answer is no, not at this time.
Longer answer: the raw C++ for .yxdb support is available on github, as it was open sourced in order to adhere to R licensing when Alteryx hooked into R. See this link where Ned Harding explains it all in his blog. So basically, everything is there for someone to build Python support by utilizing the open source C++ ... but nobody has done so just yet.

Not python specific, a YXDB to SQLite DB command line based on the C++ library open sourced by Alteryx.
Limitations:
Not a Python module: use subprocess to invoke the command, then pandas/sqlite3 to read from the SQLite database file.
Read YXDB into SQLite only: write to YXDB not implemented (altough the alteryx library would allow it)
Disclaimer: I'm the author of the fork.

Dynamically creating python class from a protobuf file at run time?

In Python, is it possible at run time to convert a Google Protocol Buffers .proto file into a python class that reads that data? Python is a very dynamic language. When you use protoc to convert a .proto file to python source code, the generated code makes a lot of use of python metaclasses, so it's already very dynamic.
Ideally, I'm thinking of something like this:
import whatever
module = whatever.load_from_file("myfile.proto")
Is this possible?
(I am new to protocol buffers, please let me know if my question makes no sense)

In theory, all the pieces exist to make this work. The Python protobuf implementation could call the C++ .proto parser library (libprotoc) as a C extension to get Descriptors, and then could feed those into the metaclasses.
However, as far as I know, no one has quite tied it altogether. (Disclaimer: My knowledge is a few years old, I may have missed a new development, but I don't see anything in the docs.)
Incidentally, Cap'n Proto's Python implementation does do what you describe, proving it is possible. But that doesn't help you if you need to work with Protobuf format.
(Disclosure: I was the author of most of Google's open source Protobuf code, and I am also the author of Cap'n Proto.)

The most elegant solution to this problem is to use init.py. See: What is __init__.py for?
You can invoke your script to generate Python protobuf classes in init.py. Upon importing, that script will be automatically invoked to generate Python protobuf classes.

solution to convert PDFs, DOCs, DOCXs into a textual format with python

I am developing a full text search engine for indexing popular binary formats. I know that there are hundereds of such questions (and solutions) already, but I found it tough to find one:
cross platform
supports DOC, DOCX and PDF formats at once
easy to use with python
can be set up in a major shared host

For PDFs, I recommend PDFminer.
Try the docx module (I have not used it myself)
I am not aware of any pure python module that can read .doc files.
There are command-line tools to extract text from .doc files: antiword and catdoc (and probably others). If the packages are installed on your shared host, you could use subprocess to shell out to these tools. Available on Windows via Cygwin.
Apache POI is a Java library that can extract text from Office documents. If your shared host has Java installed, you could write a bit of Java (or Jython) code and execute using subprocess.

If at server side you can use OpenOffice then you can use unoconv: Convert between any document format supported by OpenOffice

One possible solution is to use google documents to extract the text contents from binary .doc-files. You upload the document to google docs and then download the text contents. It is a fairly slow process, but it is the only "pure Python" solution I know of since it doesn't require any external tools except for network access. An external tool such as catdoc or antiword is a much better solution if you are allowed to install it on your host.

Textract uses the default tools for every kind of file.
https://github.com/deanmalmgren/textract

How to extract a windows cabinet file in python

Is it somehow possible to extract .cab files in python?

Not strictly answering what you asked, but if you are running on a windows platform you could spawn a process to do it for you.
Taken from Wikipedia:
Microsoft Windows provides two
command-line tools for creation and
extraction of CAB files. They are
MAKECAB.EXE (included within Windows
packages such as 'ie501sp2.exe' and
'orktools.msi'; also available from
the SDK, see below) and EXTRACT.EXE
(included on the installation CD),
respectively. Windows XP also provides
the EXPAND.EXE command.

I had the same problem last week so I implemented this in python. Comments, additions and especially pull requests welcome: https://github.com/hughsie/python-cabarchive

Oddly, the msilib can only create or append to .CAB files, but not extract them. :(
However, the hachoir parser module can apparently read & edit Cabinets. (I have not used it, though, so I couldn't tell you how fitting it is or not!)

Python lib to Read a Flash swf Format File

I'm interested in using Python to hack on the data in Flash swf files. There is good documentation available on the format of swf files, and I am considering writing my own Python lib to parse that data out using the standard Python struct lib.
Does anybody know of a Python project that already does this? I would also be interested in any available solutions that use Perl, Ruby, Haskell, etc.

Well, unless you're doing it for fun (in which case, go for it!), why not use Ming? It supposedly has python wrappers...

I found another option in SWF Tools. They provide a Python wrapper that supports generating SWF files in Python.
I'm not sure if either SWF Tools or Ming actually supports parsing in and modifying an existing swf file, however. Both seem geared more towards generating swf files from scratch.