Merge pull request #1379 from kwmsmith/remove-outdated-dynd

Remove DyND for now.

.travis.yml

@@ -65,9 +65,6 @@ install:
   # Install PyMongo
   - pip install pymongo
 
-  # Install dynd
-  - conda install -c mwiebe dynd-python
-
   # Install coveralls
   - pip install coveralls
 

README.rst

@@ -141,7 +141,7 @@ follows
 
 ::
 
-    conda install blaze spark dynd-python -c libdynd -c blaze -c anaconda-cluster -y
+    conda install blaze spark -c blaze -c anaconda-cluster -y
     conda remove odo blaze blaze-core datashape -y
 
 After running these commands, clone ``odo``, ``blaze``, and ``datashape`` from

all.recipe/meta.yaml

@@ -23,7 +23,6 @@ requirements:
 test:
   requires:
     - pytest >=2.7.0
-    - dynd-python
 
   commands:
     - py.test -vx --doctest-modules --doctest-ignore-import-errors --pyargs blaze -rsxX

blaze-expr-function-correspondence.md

@@ -0,0 +1,53 @@
+Our overall goal is to focus on Blaze from an end user's (and new user's)
+perspective.  We want to identify common usability issues, address them, and
+help make blaze a more useful technology.
+
+One way we want to improve Blaze is by opening it up and allowing it to work
+better with Python as a whole.
+
+To this end, we want to build a bridge between users' existing Python
+knowledge and Blaze.  We can use this bridge to help new users better
+understand Blaze expressions, and we can highlight the correspondences.
+
+There is an interesting correspondence between Blaze expressions and Python
+functions:
+
+Blaze | Python
+------|-------
+expression | (pure) Python function
+symbol     | unbound function arguments
+bound symbol | function closure
+`compute(expr, data)` | `func(**data)`
+
+We can take this correspondence further, and we think that would be a good thing to do:
+
+Python | Blaze
+------|-------
+function docstring | automatic blaze expression docstring, with ability for end users to 
+
+
+
+Improve Blaze design to play nicely with rest of Python, highlight correspondences:
+Blaze expressions ⇔ Python functions
+Blaze symbols ⇔ unbound function arguments
+Blaze bound symbols ⇔ closure / default argument values
+(new) automatic blaze expr docstrings ⇔ function docstrings
+(new) callable expressions ⇔ calling Python function
+(new) composing blaze expressions ⇔ composing Python functions
+(new) computational pipelines in Blaze: 
+pipeline = comp(filter, grouper, selector, summarizer)
+pipeline(data) => summarize(select(group(filter(data))))
+Blaze free to compose and optimize expressions in pipeline.
+Remove distinction between “interactive” and “non-interactive” interaction with Blaze.
+Extremely confusing to new users.  Change terminology, documentation, API naming, etc.
+Still keep bound and unbound symbols.
+Expressions can be fully unbound, partially bound, or fully bound.  Have to have a fully bound expression to actually compute a result.
+compute(expr, {mapping for remaining unbound symbols})
+make all computation explicit, but still easily accessible.
+No more implicit computation in `__repr__`.
+`compute()` changed so that it actually computes, every time.
+if we want compilation of blaze expression to SQLA query, call that something else (`compile()` maybe?).  Can be used as part of compute’s operation.
+By default, compute returns results in default backend, odos to default backend if necessary.
+all table-like things: default is Pandas; dask.dataframe if ooc.
+all array-like things: default is NumPy; dask.array if ooc.
+Overrideable behavior with kwarg

blaze/__init__.py

@@ -47,9 +47,6 @@
     from .compute.spark import *
 with ignoring(ImportError, TypeError):
     from .compute.sparksql import *
-with ignoring(ImportError):
-    from dynd import nd
-    from .compute.dynd import *
 with ignoring(ImportError):
     from .compute.h5py import *
 with ignoring(ImportError):

docs/source/dev_workflow.rst

@@ -20,7 +20,6 @@ on github.com/blaze/project-name
 *  DataShape_
 *  Odo_
 *  Dask_
-*  DyND_
 
 Bleeding edge binaries are kept up-to-date on the ``blaze`` conda channel.
 New developers likely only need to interact with one or two of these libraries so we recommend downloading everything by the conda channel and then only cloning those git repositories that you actively need::
@@ -29,7 +28,6 @@ New developers likely only need to interact with one or two of these libraries s
     git clone git://github.com/blaze/blaze.git  # only clone blaze and odo
     git clone git://github.com/blaze/odo.git  # only clone blaze and odo
 
-.. _DyND: https://github.com/blaze/dynd-python
 .. _Odo: https://github.com/blaze/odo
 .. _Dask: https://github.com/blaze/dask
 .. _Blaze: https://github.com/blaze/blaze

docs/source/index.rst

@@ -25,14 +25,16 @@ project itself.
   other data storage systems.  One Blaze query can work across data ranging
   from a CSV file to a distributed database.
 
-  Blaze presents a pleasant and familiar interface to us regardless of
-  what computational solution or database we use (e.g. Spark, Impala, SQL databases,
-  No-SQL data-stores, raw-files).  It mediates our interaction with files, data structures,
-  and databases, optimizing and translating our query as appropriate to provide a smooth and interactive
-  session.   It allows the data scientists and analyst to write their queries in a unified way that does
-  not have to change because the data is stored in another format or a different data-store.
-  It also provides a server-component that allows URIs to be used to easily serve views on data and refer to Data
-  remotely in local scripts, queries, and programs.
+  Blaze presents a pleasant and familiar interface to us regardless of what
+  computational solution or database we use (e.g. Spark, Impala, SQL databases,
+  No-SQL data-stores, raw-files).  It mediates our interaction with files, data
+  structures, and databases, optimizing and translating our query as
+  appropriate to provide a smooth and interactive session.   It allows the data
+  scientists and analyst to write their queries in a unified way that does not
+  have to change because the data is stored in another format or a different
+  data-store.  It also provides a server-component that allows URIs to be used
+  to easily serve views on data and refer to Data remotely in local scripts,
+  queries, and programs.
 
 * DataShape_: A data type system
 

docs/source/install.rst

@@ -79,7 +79,6 @@ Blaze can help you use a variety of other libraries like ``sqlalchemy`` or
 ``h5py``.  If these are installed then Blaze will use them.  Some of these are
 non-trivial to install.  We recommend installation throgh ``conda``.
 
-* dynd-python_ >= 0.6.5
 * sqlalchemy_
 * h5py_
 * spark_ >= 1.1.0
@@ -94,7 +93,6 @@ non-trivial to install.  We recommend installation throgh ``conda``.
 .. _odo: https://github.com/blaze/odo
 .. _h5py: http://docs.h5py.org/en/latest/
 .. _pytest: http://pytest.org/latest/
-.. _dynd-python: https://github.com/blaze/dynd-python
 .. _datashape: https://github.com/blaze/datashape
 .. _pandas: http://pandas.pydata.org/
 .. _cytoolz: https://github.com/pytoolz/cytoolz/

docs/source/whatsnew/0.10.0.txt

@@ -0,0 +1,38 @@
+Release |version|
+-----------------
+
+:Release: |version|
+:Date: December 17th, 2015
+
+New Expressions
+~~~~~~~~~~~~~~~
+
+Improved Expressions
+~~~~~~~~~~~~~~~~~~~~
+
+New Backends
+~~~~~~~~~~~~
+
+Improved Backends
+~~~~~~~~~~~~~~~~~
+
+Experimental Features
+~~~~~~~~~~~~~~~~~~~~~
+
+API Changes
+~~~~~~~~~~~
+
+* Due to API instability and on the recommendation of DyND developers, we
+  removed the DyND dependency temporarily (:issue:`1379`).  When DyND achieves
+  its 1.0 release, DyND will be re-incorporated into Blaze.  The existing DyND
+  support in Blaze was rudimentary and based on an egregiously outdated and
+  buggy version of DyND.  We are aware of no actual use of DyND via Blaze in
+  practice.
+
+Bug Fixes
+~~~~~~~~~
+
+
+Miscellaneous
+~~~~~~~~~~~~~
+