Merge pull request #1139 from jcb91/execute_time

[ExecuteTime] add preprocessor to execute notebook updating timing metadata

Author: jcb91

docs/source/exporting.rst

@@ -21,6 +21,12 @@ Generic documentation for preprocessors can be found at
 `nbconvert.readthedocs.io/en/latest/api/preprocessors.html <http://nbconvert.readthedocs.io/en/latest/api/preprocessors.html>`__.
 
 
+Executing and updating timing metadata
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. autoclass:: ExecuteTimePreprocessor
+
+
 Retaining Codefolding
 ^^^^^^^^^^^^^^^^^^^^^
 

src/jupyter_contrib_nbextensions/nbconvert_support/__init__.py

@@ -4,6 +4,7 @@
 
 from .collapsible_headings import ExporterCollapsibleHeadings
 from .embedhtml import EmbedHTMLExporter
+from .execute_time import ExecuteTimePreprocessor
 from .exporter_inliner import ExporterInliner
 from .nbTranslate import NotebookLangExporter
 from .pp_highlighter import HighlighterPostProcessor, HighlighterPreprocessor
@@ -16,6 +17,7 @@
     'templates_directory',
     'CodeFoldingPreprocessor',
     'EmbedHTMLExporter',
+    'ExecuteTimePreprocessor',
     'ExporterCollapsibleHeadings',
     'ExporterInliner',
     'HighlighterPostProcessor',

src/jupyter_contrib_nbextensions/nbconvert_support/execute_time.py

@@ -0,0 +1,40 @@
+"""
+Module containing a preprocessor tto execute code cells, updating time metadata
+"""
+
+from datetime import datetime
+
+from nbconvert.preprocessors.execute import ExecutePreprocessor
+
+try:
+    # notebook >= 5.0.0-rc1
+    import notebook._tz as nbtz
+except ImportError:
+    # notebook < 5.0.0-rc1
+    import notebook.services.contents.tz as nbtz
+
+
+class ExecuteTimePreprocessor(ExecutePreprocessor):
+    """
+    Executes all the cells in a notebook, updating their ExecuteTime metadata.
+    """
+
+    def run_cell(self, cell, cell_index, *args, **kwargs):
+        before = datetime.utcnow()
+        exec_reply, outs = super(ExecuteTimePreprocessor, self).run_cell(
+            cell, cell_index, *args, **kwargs)
+
+        if exec_reply.get('msg_type', '') == 'execute_reply':
+            ets = cell.setdefault('metadata', {}).setdefault('ExecuteTime', {})
+            if 'started' in exec_reply.get('metadata', {}):
+                # started value should is already a string, so don't isoformat
+                ets['start_time'] = exec_reply['metadata']['started']
+            else:
+                # attempt to fallback to datetime obj for execution request msg
+                ets['start_time'] = exec_reply.get(
+                    'parent_header', {}).get('date', before).isoformat()
+            ets['end_time'] = (
+                exec_reply.get('header', {}).get('date') or nbtz.utcnow()
+            ).isoformat()
+
+        return exec_reply, outs

tests/test_preprocessors.py

@@ -1,12 +1,16 @@
 # -*- coding: utf-8 -*-
 
+import json
 import os
+import re
 
 import nbformat.v4 as nbf
-from nbconvert import LatexExporter, RSTExporter
+from nbconvert import LatexExporter, NotebookExporter, RSTExporter
 from nbconvert.utils.pandoc import PandocMissing
 from nose.plugins.skip import SkipTest
-from nose.tools import assert_in, assert_not_in, assert_true
+from nose.tools import (
+    assert_greater_equal, assert_in, assert_not_in, assert_true,
+)
 from traitlets.config import Config
 
 
@@ -107,3 +111,38 @@ def test_preprocessor_svg2pdf():
     assert_true(pdf_existed, 'exported pdf should exist')
     assert_in('test.pdf', body,
               'exported pdf should be referenced in exported notebook')
+
+
+def _normalize_iso8601_timezone(timestamp_str):
+    # Zulu -> +00:00 offset
+    timestamp_str = re.sub(r'Z$', r'+00:00', timestamp_str)
+    # HH -> HH:00 offset
+    timestamp_str = re.sub(r'([+-]\d\d)$', r'\1:00', timestamp_str)
+    # HHMM -> HH:MM offset
+    timestamp_str = re.sub(r'([+-]\d\d):?(\d\d)$', r'\1:\2', timestamp_str)
+    return timestamp_str
+
+
+def test_preprocessor_execute_time():
+    """Test ExecuteTime preprocessor."""
+    # check import shortcut
+    from jupyter_contrib_nbextensions.nbconvert_support import ExecuteTimePreprocessor  # noqa E501
+    notebook_node = nbf.new_notebook(cells=[
+        nbf.new_code_cell(source="a = 'world'"),
+        nbf.new_code_cell(source="import time\ntime.sleep(2)"),
+    ])
+    body, resources = export_through_preprocessor(
+        notebook_node, ExecuteTimePreprocessor, NotebookExporter, 'ipynb')
+    cells = json.loads(body)['cells']
+    for cell in cells:
+        if cell['cell_type'] != 'code':
+            assert_not_in('ExecuteTime', cell['metadata'])
+        else:
+            assert_in('ExecuteTime', cell['metadata'])
+            etmd = cell['metadata']['ExecuteTime']
+            assert_in('start_time', etmd)
+            assert_in('end_time', etmd)
+            assert_greater_equal(
+                _normalize_iso8601_timezone(etmd['end_time']),
+                _normalize_iso8601_timezone(etmd['start_time']),
+                'end_time should not be before start_time')