eriknw / toolz / 93

Builds Branch Commit Type Ran Committer Via Coverage
93 juxt_tuple Change `juxt` to return a tuple as discussed in #174 push eriknw travis-ci pending completion   set done
92 faster_groupby Choose a version of `groupby` that is easier to understand. This is `groupby_alt` from the original commit comment in this branch. See discussion at #179. This version performs very well as groups become larger. The previous implementation perf... push eriknw travis-ci pending completion   set done
91 faster_groupby No need to import `compatibility.iteritems` in `itertoolz` push eriknw travis-ci pending completion   set done
90 faster_groupby Add a new implementation of `groupby` that uses two dicts. The previous commit was wrong. It is slower than the commit before. There was a mistake in the code used for benchmarks. I believe the new implementation has optimal performance as grou... push eriknw travis-ci pending completion   set done
89 faster_groupby Small tweak to `groupby` to improve it for "real-world" data Benchmarks have a tendency to use data that is pathologically perfect. In such pathological cases, it was unclear whether the current or previous implementation would be preferred. How... push eriknw travis-ci pending completion   set done
88 faster_groupby Add useful code comments to groupby including pythonic implementation push eriknw travis-ci pending completion   set done
87 faster_groupby Faster groupby! Issue #178 impressed upon me just how costly attribute resolution can be. In this case, `groupby` was made faster by avoiding resolving the attribute `list.append`. This implementation is also more memory efficient than the curr... push eriknw travis-ci pending completion   set done
86 faster_unique Faster unique, isdistinct, merge_sorted, and sliding_window. The `key` keyword argument to `unique` was changed from `identity` to `None`. This better matches API elsewhere, and lets us remove `identity` from being redefined in `itertoolz`, which... push eriknw travis-ci pending completion   set done
85 faster_remove Faster `remove` by using `filterfalse`, which was added to `compatibility` Also, a typo (it lied!) was fixed in the documentation of `remove` By using `filterfalse`, we modify the behavior of `remove` when `None` is used as the predicate. Previ... push eriknw travis-ci pending completion   set done
84 equality_hash_key Refinements to `EqualityHashKey` including requiring a `key` argument. The canonical purpose of `key` is to calculate a partial key hash that is good enough. This is *much* better than using the default hash value. `'__default__hashkey__'` is u... push eriknw travis-ci pending completion   set done
83 equality_hash_key Fix typo push eriknw travis-ci pending completion   set done
82 equality_hash_key oops python3 push eriknw travis-ci pending completion   set done
81 equality_hash_key Add `toolz.sandbox.EqualityHashKey` as discussed in #166 `EqualityHashKey` creates a hash key that uses equality comparisons between items, which may be used to create hash keys for otherwise unhashable types. The tradeoffs for using this is dis... push eriknw travis-ci pending completion   set done
80 curry_makeover Pickle `curried_func.__dict__` too. push eriknw travis-ci pending completion   set done
79 curry_makeover Calculate hash in `__hash__` for curry, not in `__init__`. Also, don't return `self` from `__call__`; return a new `curry` object instead. push eriknw travis-ci pending completion   set done