8000 Tags · google/pyglove · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Tags: google/pyglove

Tags

v0.4.4

Toggle v0.4.4's commit message
Support markdown quotes on the repr/str of Formattable objects.

This allows users to use `pg.str_format` and `pg.repr_format` context manager to add markdown quotes to the repr/str of symbolic objects, which serves the scenario when embedding objects within natural language string.

PiperOrigin-RevId: 595679721

v0.4.3

Toggle v0.4.3's commit message
Support the creation of `pg.typing.Enum` from `typing.Literal` annota…

…tion.

PiperOrigin-RevId: 5651681
8000
39

v0.4.2

Toggle v0.4.2's commit message
Requiring `@pg.explicit_method_override` for overriding PyGlove manag…

…ed methods.

This prevents users from accidentally overriding PyGlove managed methods such as `pg.Object.__init__`, with user-friendly error message.

PiperOrigin-RevId: 551634474

v0.4.1

Toggle v0.4.1's commit message
`pg.typing` to support `typing.Generic`.

PiperOrigin-RevId: 550112605

v0.4.0

Toggle v0.4.0's commit message
Bump version to 0.4.0 for a major release.

PiperOrigin-RevId: 547832331

v0.3.0

Toggle v0.3.0's commit message
Smart cross references for PyGlove symbols in Sphinx autodoc.

This allows developers to refer to PyGlove symbols in Sphinx documentation in the same way as we write code, e.g. :class:`pg.DNA` or :class:`pg.geno.DNA`, both will output a link with the same canonical symbol name `pg.DNA`. As a result, all mentions of the same symbol will have consistent name across the entire document corpus.

PiperOrigin-RevId: 517022666

v0.2.1

Toggle v0.2.1's commit message
Fix bad indent for class_schema_test.py

PiperOrigin-RevId: 502696111

v0.2.0

Toggle v0.2.0's commit message
Rename 'math' and 'random' under `pg.scalars` to avoid name clash.

PiperOrigin-RevId: 480936714

v0.1.1

Toggle v0.1.1's commit message
Dynamic evaluation to support divide-and-conquer on a search space.

Example:

```
def fun():
  return pg.oneof([1, 2], hints='ssd1') + pg.oneof([3, 4], hints='ssd2')

ssd1 = pg.hyper.DynamicEvaluationContext(where=lambda x: x.hints == 'ssd1')
ssd2 = pg.hyper.DynamicEvaluationContext(where=lambda x: x.hints == 'ssd2')

# Partitioning the search space into ssd1 and ssd2.
with ssd1.collect():
  with ssd2.collect():
    fun()

# Nested search.
for ex1, f1 in pg.sample(ssd1, algorithm1):
  rs = []
  for ex2, f2 in pg.sample(ssd2, algorithm2):
    with ex1():
      with ex2():
        r = fun()
        f2(r)
        rs.append(r)
  f1(sum(rs))
```

PiperOrigin-RevId: 480115321
0