Status?

[akira]

Hi! Is Doqu still being developed? When would there be new docs?

Thanks

[Andy Mikhaylenko]

Hi,

thanks for your interest.

> Hi! Is Doqu still being developed?

Sure. The development has slowed down a bit because a) I haven't got
much time recently, and b) at its current state Doqu is already more or
less stable (in terms of API and reliability) and I'm successfully using
it in some of my projects.

The latest changes are not fully published yet. A brief summary:

1. added a unified test suite for all backends, so it is guaranteed that
the storage/query API is consistent with all supported databases (with a
couple of known issues, see tests);
2. the Document class is now optional (plain dicts are OK for schemaless
data).
3. We have discussed performance issues. I've written a small testing
suite and made some optimizations; there's still a lot to be done in
this vein.

(By the way, the Document class is currently not only very powerful but
also very slow; use plain dicts if performance is more important than
bidirectional validation. Or drop in your own class, it's easy to do.)

> When would there be new docs?

I really think that documentation is almost as important as code, so
current state of Doqu docs is a pain. However, the tests and docstrings
should tell a lot about the API. The basic use cases are the same as in
the old docs but the library has become much more flexible and the docs
should reflect this.

So, the rough roadmap is:

1. Finish refactoring to achieve loose coupling;
2. Write tests for the document/query goodies (validation, etc.);
3. Write a performance-centric test suite (for comparative regression
testing: assert a certain overhead against underlying libs to track the
effect of each commit on both functionality and performance);
4. Write a new tutorial;
5. Add test cases (e.g. query operations) to docs as a reference.

If you'd like to get some help or share ideas, you are welcome. For
instance, it may be easier for me to help you solve a specific problem
and then copy the discussion to the docs than to write the whole thing
from scratch without any real questions.

Cheers,
Andy