Add jupyter configuration files to export .jl files

Add Jupyter configs to automatically export a script file whenever a
Jupyter Notebook file (.ipynb) in the repo is saved. This makes diffing
changes in git significantly easier.

The files are exported to a top-level directory called .nbexports, which
mirrors the filepaths of the real files. (Note, when scripts are renamed
/ deleted, it will require manual cleanup from the .nbexports directory)

Author: NHDaly

jupyter_notebook_config.py

@@ -0,0 +1,42 @@
+import io
+import os
+
+dir_path = os.path.dirname(os.path.realpath(__file__))
+nbexports_path = os.path.join(dir_path, ".nbexports")
+
+from notebook.utils import to_api_path
+
+_script_exporter = None
+
+def script_post_save(model, os_path, contents_manager, **kwargs):
+    if model['type'] != 'notebook':
+        return
+
+    from nbconvert.exporters.script import ScriptExporter
+    from nbconvert.exporters.html import HTMLExporter
+
+    global _script_exporter
+    if _script_exporter is None:
+        _script_exporter = ScriptExporter(parent=contents_manager)
+        _script_exporter.template_file = os.path.join(dir_path, 'jupyter_script_export_template.tpl')
+
+    export_script(_script_exporter,   model, os_path, contents_manager, **kwargs)
+
+def export_script(exporter, model, os_path, contents_manager, **kwargs):
+    """convert notebooks to Python script after save with nbconvert
+    replaces `ipython notebook --script`
+    """
+    base, ext = os.path.splitext(os_path)
+    script, resources = exporter.from_filename(os_path)
+    script_fname = base + resources.get('output_extension', '.txt')
+    script_repopath = to_api_path(script_fname, contents_manager.root_dir)
+    log = contents_manager.log
+    script_fullpath = os.path.join(dir_path, ".nbexports", script_repopath)
+    os.makedirs(os.path.dirname(script_fullpath), exist_ok=True)
+    log.info("Saving script /%s", script_fullpath)
+    with io.open(script_fullpath, 'w', encoding='utf-8') as f:
+        f.write(script)
+
+
+c.FileContentsManager.post_save_hook = script_post_save
+

jupyter_script_export_template.tpl

@@ -0,0 +1,63 @@
+{# Lines inside these brackets are comments #}
+{#- Brackets with a `-` mean to skip whitespace before or after the block. -#}
+
+{#-
+ # This file defines a Jinja Template for converting .ipynb files into scripts
+ # for either delve or julia. We need this because the default script exporter
+ # doesn't render Markdown for any languages except Python.
+ # Exporting the markdown makes github reviews of .ipynb files easier.
+ # This template is invoked by our custom `jupyter_notebook_config.py`. You can
+ # read more about the Jinja template specification here:
+ #   http://jinja.pocoo.org/docs/2.10/templates/#comments
+ # and here:
+ #   https://nbconvert.readthedocs.io/en/latest/customizing.html
+ # And the filter functions used in this file are defined here:
+ #   https://github.com/pallets/jinja/blob/master/jinja2/filters.py
+ # and here:
+ #   https://github.com/jupyter/nbconvert/blob/master/nbconvert/filters/strings.py
+ -#}
+
+{#- ---------
+ # Lines up here, before the `extends` section, go at the top of the file, before any other
+ # content from the notebook itself.
+ #----------- -#}
+
+{%- if 'name' in nb.metadata.get('kernelspec', {}) and
+    nb.metadata.kernelspec.name == 'julia' -%}
+# This file was generated from a Julia language jupyter notebook.
+{% endif -%}
+
+{% extends 'script.tpl'%}
+
+{% block markdowncell %}
+  {#- Turn the contents of the markdown cell into a wrapped comment block, and trim empty lines. -#}
+
+  {#-
+   # NOTE: We used `kernelspec.name` not `language_info.name`, for reasons specific to
+   # our custom jupyter kernel. I think `language_info.name` might be more robust?
+   -#}
+  {%- if 'name' in nb.metadata.get('kernelspec', {}) and
+    nb.metadata.kernelspec.name == 'julia' -%}
+    {%- set commentprefix = '# ' -%}
+  {#-
+   # Add other languages here as if-else block, e.g. C++ would use '// '
+   -#}
+  {%- else -%}
+    {#- Assume python by default -#}
+    {%- set commentprefix = '# ' -%}
+  {%- endif -%}
+
+  {%- set commentlen = 92-(commentprefix|length) -%}
+  {{- '\n' -}}
+  {{- commentprefix ~ '-' * commentlen -}}
+  {{- '\n' -}}
+
+  {#- Turn the contents of the markdown cell into a wrapped comment block, and trim empty lines. -#}
+  {#- Note: `comment_lines` and `wrap_text` are defined in nbconvert/filters/strings.py -#}
+  {{- cell.source | wrap_text(width=commentlen) | comment_lines(prefix=commentprefix) | replace(commentprefix~"\n", commentprefix|trim ~ "\n") -}}
+
+  {{- '\n' -}}
+  {{- commentprefix ~ '-' * commentlen -}}
+  {{- '\n' -}}
+
+{% endblock markdowncell %}