New release for pyfolio

[eigenfoo]

It looks like enough has changed on pyfolio to merit v0.8.0 (with risk and performance attribution, I think we might even want to call it v1.0.0!)

We are also seeing an increasing number of users running into bugs that have already been patched: upping the version would make it easier to diagnose errors (both ours, and open source users).

To-do:

• Performance attribution tear sheet ENH: add perf attribution + summary table #417 (still WIP, will definitely want to wait for this)
• Resolve any dependency issues with zipline/alphalens/empyrical
• Do we want to outsource Fama-French functionality from pyfolio to empyrical? #56 on empyrical (Answer: yes. PR waiting: remove old data loaders from utils, and migrate to empyrical #416)
• Do we want to include returns decomposition from alphalens into pyfolio? #183 on alphalens

@twiecki @gusgordon @richafrank @Jstauth thoughts? If this sounds good I'll organize a milestone :)

New features

• Risk tear sheet: added a new tear sheet to analyze risk exposures to common factors, sector, market cap and illiquidity (Risk tear sheet #391)
• Simple tear sheet: added a new tear sheet that presents only the most important plots in the full tear sheet, for a quick general overview of a portfolio's performance (Simple Tearsheet #389)
• Volatility plot: added a rolling annual volatility plot in the returns tear sheet (fixes #278 #390)

Bugfixes

• Fixed bug regarding Yahoo API and pandas data reader (Fix #385 among other things #395)
• Removed information_ratio to remain compatible with empyrical
• Fama-French rolling multivariate regression bug fix (Fama-French multivariate regression #406)

Maintenance

• Updated and improved jupyter notebook documentation (Fix example notebooks #411)

[twiecki]

Yes, should absolutely cut a new release! Want to do a PR that updates the version number and has a release announcement with a list of new features (similar to what you have here)?

[eigenfoo]

I'll draft a release and organize a milestone so we know what PRs to wait for and what not to. A milestone will be better than an issue :)

[eigenfoo]

I just remembered @twiecki I had drafted a new release announcement on github, but since my access has been revoked I assume its lost. You can just use my comment above to draft a new one.

[twiecki]

Nothing should get lost. It's either on a branch you pushed or on a local branch of yours. Where did you keep it?

[eigenfoo]

As a draft release: click "10 releases" on the info bar on the pyfolio github home page.

[eigenfoo]

@twiecki @gusgordon once @vikram-narayan's perf attributions are done we can cut a new version of pyfolio, everything else is merged.

[twiecki]

We can probably already do that now given the changes and refactorings we have.

[eigenfoo]

If you think that would be a good idea! 🎉

[twiecki]

@eigenfoo Want to do a PR with the release announcements and version bump to 0.8.0?

[eigenfoo]

Will do!

[bonesmccoy]

Hi, isn't the changes of requirements breaking the requirements of zipline 1.1.1

zipline 1.1.1 : pandas==0.18.1
pyfolio 0.8.0: pandas>=0.19.0

zipline 1.1.1: empyrical==0.2.2
pyfolio 0.8.0: empyrical>=0.3.2

Source:
Pyfolio: https://github.com/quantopian/pyfolio/blob/v0.8.0/setup.py
Zipline: https://github.com/quantopian/zipline/blob/1.1.1/etc/requirements.txt

you may say: you don't need ZL with pyfolio.
Yes but when I'm trying to open a zipline result pickle file, if I don't have ZL Pyfolio is not able to unserialize the ZL Objects.

Am I doing something wrong?
Any suggestion?

[twiecki]

@bonesmccoy Hm, that's indeed a pickle (no pun intended). The empyrical version mismatch should be fixed easily, the pandas one might be more tricky. CC @richafrank

[twiecki]

We can probalby make pyfolio pandas 0.18.1 compat without too much effort (if it isn't already).

[richafrank]

Yup, we'll be fixing the pandas issue soon, unless you get to it first.

zipline doesn't read the reqs.txt literally and actually supports empyrical>=2.2, so I don't think that's an issue here, unless you specifically want to use empyrical<0.3.2.

[eigenfoo]

Closed by #552