Update binder config (#129)

* Fixing requirements for binder build

* Remove (hopefully) unnecessary dependency

* Add big packages back in

* Add BERT download to image prep

* Remove unnecessary dependencies from conda config

* Add missing import

* Update instructions for setting up credentials on binder/colab/watson studio

* Adjust kernel settings and rerun

Author: frreiss

binder/download_bert.py

@@ -0,0 +1,11 @@
+
+# download_bert.py
+#
+# Post-build script to download and cache a BERT model so that
+# Model_Training_with_BERT.ipynb won't run of disk space.
+
+import transformers
+
+bert_model_name = "dslim/bert-base-NER"
+bert = transformers.BertModel.from_pretrained(bert_model_name)
+

binder/postBuild

@@ -3,9 +3,27 @@
 # Log commands
 set -ex
 
+# Not sure why we need this here.
+pip install -r requirements.txt
+
 # Install additional packages beyond those in requirements.txt so that the
 # JupyterLab instance will be able to run all of the NLP library integrations.
 pip install -r config/dev_reqs.txt
 pip install -r config/jupyter_reqs.txt
 
+# Install large packages during image build for now.
+# Later we may want to have the notebooks install these on demand if they
+# detect they are running in a Binder environment.
+# Note that transformers will start up in a bad state if it detects no PyTorch
+# and will stay in that state until Python is restarted.
+pip install -r config/big_reqs.txt
+
+# Download the language model needed by Integrate_NLP_Libraries.ipynb
+# We can't download this package from within the notebook unless the user
+# restarts the kernel after the download.
+python -m spacy download en_core_web_sm
+
+# Also download the BERT model used by Model_Training_with_BERT.ipynb so that
+# the production environment won't run out of disk space.
+python ./binder/download_bert.py
 

config/big_reqs.txt

@@ -0,0 +1,6 @@
+# big_reqs.txt
+# Large packages that are required by a small fraction of our demo
+# material.
+scikit-learn
+torch
+

config/dev_env.yml

@@ -3,14 +3,5 @@
 channels:
   - defaults
 dependencies:
- - tensorflow
- - pandas
- - regex
- - matplotlib
- - cython 
- - grpcio-tools
- - pytorch
- - black
- - scikit-learn 
 
 

config/dev_reqs.txt

@@ -8,7 +8,6 @@ pyyaml
 transformers>=3.0.0
 # SpaCy models aren't stable across point releases
 spacy==2.3.2
-fastparquet
 sphinx
 sphinxcontrib-apidoc
 ipywidgets

config/jupyter_reqs.txt

@@ -5,7 +5,5 @@ jupyterlab
 nodejs
 xeus-python
 ipywidgets
-scikit-learn
-torch
 matplotlib
 

env.sh

@@ -80,6 +80,9 @@ pip install -r config/dev_reqs.txt
 # Additional layer of pip-installed stuff for running notebooks
 pip install -r config/jupyter_reqs.txt
 
+# Additional large packages needed by only some of the demos
+pip install -r config/big_reqs.txt
+
 ################################################################################
 # Least-preferred install method: Custom
 

notebooks/Analyze_Text.ipynb

@@ -96,7 +96,7 @@
     "  jupyter lab\n",
     "```\n",
     "\n",
-    "Alternately, you can modify the code below to set the variables `api_key` and `service_url` directly.\n",
+    "Alternately, you can uncomment the first two lines of code below to set the `IBM_API_KEY` and `IBM_SERVICE_URL` environment variables directly.\n",
     "**Be careful not to store your API key in any publicly-accessible location!**"
    ]
   },
@@ -106,6 +106,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "# If you need to embed your credentials inline, uncomment the following two lines and\n",
+    "# paste your credentials in the indicated locations.\n",
+    "# os.environ[\"IBM_API_KEY\"] = \"<API key goes here>\"\n",
+    "# os.environ[\"IBM_SERVICE_URL\"] = \"<Service URL goes here>\"\n",
+    "\n",
     "# Retrieve the API key for your Watson NLU service instance\n",
     "if \"IBM_API_KEY\" not in os.environ:\n",
     "    raise ValueError(\"Expected Watson NLU api key in the environment variable 'IBM_API_KEY'\")\n",
@@ -138,7 +143,7 @@
     {
      "data": {
       "text/plain": [
-       "<ibm_watson.natural_language_understanding_v1.NaturalLanguageUnderstandingV1 at 0x7fc092cc1e90>"
+       "<ibm_watson.natural_language_understanding_v1.NaturalLanguageUnderstandingV1 at 0x7ffa4f50ed10>"
       ]
      },
      "execution_count": 3,
@@ -2857,7 +2862,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.7.7"
+   "version": "3.7.9"
   }
  },
  "nbformat": 4,