feature request: unicode in source

[tpapp]

Introduction

Some languages now support Unicode (mostly UTF8) for writing source code. It would be great if one could also use Unicode in Stan source. (Note that comments in UTF8, or any superset that embeds ASCII, are already supported in the sense the parser just ignores them.)

Broadly, there are two possible levels of support:

1. in variable and function names (eg ϕ), and
2. in operators (eg ≤), which provide synonyms for existing ones (eg <=)

Example

This is how the 8 schools example would look like in unicode:

data {
  int<lower=0> J;             // number of schools
  real y[J];                  // estimated treatment effect (school j)
  real<lower=0> σ[J];         // std err of effect estimate (school j)
}
parameters {
  real μ;
  real θ[J];
  real<lower=0> τ;
}
model {
  θ ~ normal(μ, τ); 
  y ~ normal(θ, σ);
}

Possible benefits

1. more compact source code
2. better mapping to equations in papers

Possible downsides

1. editor/entry support
2. font support
3. possibly corrupted files

The first two are mitigated by the fact that ASCII is a subset of UTF8, so using the feature is optional.

UTF8 support in various languages which have interfaces for Stan

language	literals	identifiers	operators	would UTF8 variables work for interfacing with Stan?
R	yes	yes	no	yes
Python	yes	only from version 3	no	yes, even in Python 2, as they are used as literal keys
Julia	yes	yes	yes	yes
Matlab	yes	yes, but needs to be enabled	no	yes
Stata	yes	yes, from version 14	no	probably?

Editor support

Emacs

See this list for various UTF8 implementations using autocomplete, company-mode, and quail.

See also

• discussion on stan-dev
• detailed description for Julia

[ariddell]

Python 2 handles UTF-8 fine. (It's just not the default representation.)

[tpapp]

@ariddell: Can you provide a link with an example/description? Then I could update the list.

[bgoodri]

RStudio should work https://support.rstudio.com/hc/en-us/articles/200532197-Character-Encoding

[ariddell]

@tpapp https://docs.python.org/2/howto/unicode.html Python 3 uses Unicode by default. In Python 2 you need to be explicit about it. In both cases you can have unicode strings in source code.

[tpapp]

@ariddell: in the page you link I could not find an example with unicode identifiers (only strings, literals, filenames, etc).

[ariddell]

Here's a link to the section: https://docs.python.org/2/howto/unicode.html#unicode-literals-in-python-source-code

[bob-carpenter]

Those are unicode literals, not unicode identifiers.

Can you have

éø = 10

where you assign to a unicode literal? Or dictionaries with unicode keys?

• Bob

On May 11, 2016, at 1:24 PM, Allen Riddell [email protected] wrote:

Here's a link to the exact section: https://docs.python.org/2/howto/unicode.html#unicode-literals-in-python-source-code

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub

[ariddell]

You're right. In Python 2.7 you can't have unicode variables. In Python 3 you can. But why does that matter? We only need unicode in the Stan program code. (Parameter lookups aren't affected since keys are and always were strings.)

[bob-carpenter]

It looks like you use a dictionary structure for variable names.

schools_dat = { 'J': 8, 'y': [28, 8, -3, 7, -1, 1, 18, 12], 'sigma': [15, 10, 16, 11, 9, 11, 10, 18] }

Can the keys be unicode?

RStan can read data values out of the environment if they're named after variables in the Stan program. And it can attach the resulting draws as variables in the environment.

• Bob

On May 11, 2016, at 4:07 PM, Allen Riddell [email protected] wrote:

You're right. In Python 2.7 you can't have unicode variables. In Python 3 you can. But why does that matter? We only need unicode in the Stan program code. (Parameter lookups aren't affected since keys are and always were strings.)

— You are receiving this because you commented. Reply to this email directly or view it on GitHub

[ariddell]

Python 2 has no problems with unicode dictionary keys. In fact, it can have unicode variables in the environment but you have to reference them via strings indirectly. For example, this works in Python 2:

>>> locals()[u'é'] = 9
>>> locals()[u'é']

(locals is something like baseenv or .GlobalEnv in R)

Bref, there is nothing Python 2 can't do that's relevant to supporting unicode in Stan code. The table above is inaccurate.

On 05/11, Bob Carpenter wrote:

It looks like you use a dictionary structure for variable names.

schools_dat = { 'J': 8, 'y': [28, 8, -3, 7, -1, 1, 18, 12], 'sigma': [15, 10, 16, 11, 9, 11, 10, 18] }

Can the keys be unicode?

RStan can read data values out of the environment if they're named after variables in the Stan program. And it can attach the resulting draws as variables in the environment.

• Bob

On May 11, 2016, at 4:07 PM, Allen Riddell [email protected] wrote:

You're right. In Python 2.7 you can't have unicode variables. In Python 3 you can. But why does that matter? We only need unicode in the Stan program code. (Parameter lookups aren't affected since keys are and always were strings.)

— You are receiving this because you commented. Reply to this email directly or view it on GitHub

---

You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/stan-dev/stanc3/issues/1406

[bob-carpenter]

On May 12, 2016, at 8:20 AM, Allen Riddell [email protected] wrote:

Python 2 has no problems with unicode dictionary keys. In fact, it can have unicode variables in the environment but you have to reference them via strings indirectly. For example, this works in Python 2:

>>> locals()[u'é'] = 9
>>> locals()[u'é']

(locals is something like baseenv or .GlobalEnv in R)

Bref, there is nothing Python 2 can't do that's relevant to supporting unicode in Stan code. The table above is inaccurate.

You should have edit permission on the issues.

• Bob

[ariddell]

I was just recording my thought on the matter. I appreciate @tpapp putting work into drafting the issue text and would prefer to leave any edits to him.

[tpapp]

@ariddell: The table was accurate, but since not all Stan interfaces work the way that R/Julia does, I extended it with the information that is probably most relevant: whether the interfaces, in the way they currently operate, would support UTF8 variables for (1) passing data to Stan and (2) extracting MCMC results. Thanks for pointing this out, this is much more important than the details of UTF8 support in those languages per se.

Not being a STATA user, I am reluctant to make a definitive statement about it. If someone could help with that it would be great.

[ariddell]

One use of unicode in Stan Program code which should definitely be supported is in comments. Leaving code comments in one's native language is fairly routine in Python/Java/etc. We should at least support that in Stan.

[bob-carpenter]

Unicode in comments is OK now.

• Bob

On May 14, 2016, at 7:28 PM, Allen Riddell [email protected] wrote:

One use of unicode in Stan Program code which should definitely be supported is in comments. Leaving code comments in one's native language is fairly routine in Python/Java/etc. We should at least support that in Stan.

— You are receiving this because you commented. Reply to this email directly or view it on GitHub

[tpapp]

Indeed UTF8 comments work fine, and I have been using them for a while. Made a clarification in the issue.

[ariddell]

UTF8 comments aren't supported in PyStan right now (non-ASCII characters will generate an error). I'll fix this. stan-dev/pystan#223