Subscriptions (#2)

* get deps up-to-date, add graphql-ws

* install subscriptions

* did someone say subscriptions?

* blacken

* sure, let's call it a version

* subscribe to something useful

Author: bollwyvl

.gitignore

@@ -5,11 +5,9 @@ _scripts/
 *_files/
 *.bundle.*
 *.egg-info/
-*.html
 *.log
 *.tar.gz
 envs/
 lib/
 Untitled*.ipynb
 static/
-anaconda-project.yml

CHANGELOG.md

@@ -0,0 +1,9 @@
+# CHANGELOG
+
+## Unreleased
+
+### 0.2.0
+- Subscriptions (in server and GraphiQL)
+
+### 0.1.0
+- Basic Capability with contents manager

anaconda-project.yml

@@ -0,0 +1,41 @@
+name: jupyter-graphql-dev
+
+commands:
+  lab:
+    unix: jupyter lab --no-browser --debug
+  setup:
+    unix: pip install -e . --ignore-installed --no-deps
+  black:
+    unix: black src/py setup.py
+  atom:
+    unix: atom .
+  static:
+    unix: python -m jupyter_graphql.fetch_static
+
+env_specs:
+  default:
+    platforms:
+    - linux-64
+    - osx-64
+    - win-64
+    inherit_from:
+    - jupyter-graphql-dev
+    packages:
+    - black
+    - flake8
+    - beautysh
+  jupyter-graphql-dev:
+    packages:
+    - gql
+    - graphene
+    - iso8601
+    - jupyterlab >=0.35,<0.36
+    - python >=3.6,<3.7
+    - requests
+    - werkzeug
+    - pip:
+      - graphql-ws
+      - graphene-tornado
+    channels:
+    - conda-forge
+    - defaults

environment.yml

@@ -5,16 +5,13 @@ channels:
   - defaults
 
 dependencies:
-  - aniso8601
+  - gql
+  - graphene
+  - iso8601
   - jupyterlab >=0.35,<0.36
-  - pip
-  - promise
   - python >=3.6,<3.7
   - requests
-  - rx
   - werkzeug
   - pip:
-    - gql
-    - graphene
-    - graphql-core
-    - graphql-relay
+    - graphql-ws
+    - graphene-tornado

notebooks/gql.ipynb

@@ -16,19 +16,26 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import requests  # should investigate writing a tornado transport\n",
+    "from pprint import pprint\n",
     "from getpass import getpass\n",
     "from gql import Client, gql\n",
     "from gql.transport.requests import RequestsHTTPTransport\n",
-    "from IPython.display import JSON\n",
-    "from pprint import pprint"
+    "from IPython.display import JSON, IFrame\n",
+    "import requests  # should investigate writing a tornado transport"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "You'll need your `jupyter notebook` or `jupyter lab` token (view source, look for `\"token\"`)"
+    "## Client!"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "This has to know where you are. For example for `http://localhost:8888/lab`:"
    ]
   },
   {
@@ -37,14 +44,14 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "token = getpass()"
+    "URL = \"http://localhost:8888/graphql\""
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Update to suit!"
+    "Since this is a _kernel_ talking back to the notebook _server_, you'll need your `jupyter notebook` or `jupyter lab` token (view source, look for `\"token\"`)"
    ]
   },
   {
@@ -53,14 +60,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "URL = \"http://localhost:8888/graphql\""
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "Make a client!"
+    "token = getpass()"
    ]
   },
   {
@@ -79,7 +79,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Make a query!"
+    "## Query!"
    ]
   },
   {
@@ -88,11 +88,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "query = gql('''\n",
-    "query {\n",
+    "query = \"\"\"{\n",
     "  contents(path: \"notebooks/gql.ipynb\") {\n",
+    "    path\n",
+    "    last_modified\n",
     "    ... on NotebookContents {\n",
-    "      path\n",
     "      content {\n",
     "        nbformat\n",
     "        nbformat_minor\n",
@@ -107,8 +107,9 @@
     "    }\n",
     "  }\n",
     "}\n",
-    "''')\n",
-    "query"
+    "\"\"\"\n",
+    "query_gql = gql(query)\n",
+    "query_gql"
    ]
   },
   {
@@ -124,15 +125,50 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "result = client.execute(query)\n",
+    "result = client.execute(query_gql)\n",
     "JSON(result)"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Where can you go from here?"
+    "Where can you go from here? "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Subscribe!\n",
+    "With a little work up-front and even less work at query time, the same types from above can be used to power live _subscriptions_. Right now, only contents are available, but many things in the notebook server and broader ecosystem could become \"live\"."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "subscription = f\"subscription {query}\"\n",
+    "print(subscription)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Go ahead and paste that in the iframe below and hit (▷)!\n",
+    "> TODO: fix query param parsing!"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "IFrame(URL, width=\"100%\", height=\"400px\")"
    ]
   }
  ],
@@ -152,7 +188,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.6"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,

postBuild

@@ -25,8 +25,12 @@ classifiers =
 
 [options]
 install_requires =
-    jupyterlab
     graphene-tornado
+    graphql-ws
+    iso8601
+    notebook
+    werkzeug
+
 package_dir =
     = src/py
 packages = find:

setup.cfg

@@ -1,11 +1,10 @@
 from pathlib import Path
 from notebook.base.handlers import FileFindHandler
 
-
 from notebook.utils import url_path_join as ujoin
 
-from .handlers import GraphQLHandler
-
+from .subscriptions import TornadoSubscriptionServer
+from .handlers import GraphQLHandler, SubscriptionHandler
 from .schema import schema
 
 
@@ -18,6 +17,8 @@ def load_jupyter_server_extension(app):
     app.log.info("[graphql] initializing")
     web_app = app.web_app
 
+    subscription_server = TornadoSubscriptionServer(schema)
+
     # add our templates
     web_app.settings["jinja2_env"].loader.searchpath += [TEMPLATES]
 
@@ -34,12 +35,12 @@ def app_middleware(next, root, info, **args):
             (
                 base(),
                 GraphQLHandler,
-                dict(
-                    schema=schema,
-                    graphiql=True,
-                    nb_app=app,
-                    middleware=[app_middleware],
-                ),
+                dict(schema=schema, graphiql=True, middleware=[app_middleware]),
+            ),
+            (
+                base("subscriptions"),
+                SubscriptionHandler,
+                dict(subscription_server=subscription_server, app=app),
             ),
             # serve the graphiql assets
             (base("static", "(.*)"), FileFindHandler, dict(path=[STATIC])),

src/py/jupyter_graphql/__init__.py

@@ -1,28 +1,44 @@
 from urllib.request import urlretrieve
 from urllib.parse import urlparse
+import sys
 
 from . import STATIC
 
 # Download the file from `url` and save it locally under `file_name`:
 
-ASSETS = [
+JSDELIVR_ASSETS = [
     "https://cdn.jsdelivr.net/npm/[email protected]/graphiql.css",
     "https://cdn.jsdelivr.net/npm/[email protected]/fetch.min.js",
     "https://cdn.jsdelivr.net/npm/[email protected]/umd/react.production.min.js",
     "https://cdn.jsdelivr.net/npm/[email protected]/umd/react-dom.production.min.js",
     "https://cdn.jsdelivr.net/npm/[email protected]/graphiql.min.js",
+    # "https://cdn.jsdelivr.net/npm/[email protected]/browser/client.js",
+    # "https://cdn.jsdelivr.net/npm/[email protected]/dist/fetcher.js",
+]
+
+UNPKG_ASSETS = [
+    "https://unpkg.com/[email protected]/browser/client.js",
+    "https://unpkg.com/[email protected]/browser/client.js",
 ]
 
 
-def fetch_static():
-    for url in ASSETS:
-        out = (STATIC / urlparse(url).path[1:]).resolve()
-        if not out.exists():
+def fetch_assets(assets, prefix=None, force=False):
+    for url in assets:
+        out = STATIC
+        if prefix:
+            out = STATIC / prefix
+        out = (out / urlparse(url).path[1:]).resolve()
+        if force or not out.exists():
             out.parent.mkdir(parents=True, exist_ok=True)
             out.write_text("")
-            print("fetching", url, "to", out)
+            print(f"fetching\n\t- {url}\n\t> {out.relative_to(STATIC)}")
             urlretrieve(url, out)
 
 
+def fetch_static(force=False):
+    fetch_assets(JSDELIVR_ASSETS, force=force)
+    fetch_assets(UNPKG_ASSETS, "npm", force=force)
+
+
 if __name__ == "__main__":
-    fetch_static()
+    fetch_static(force="--force" in sys.argv)