Fix SQL multimodal example

superduper-io · Nov 29, 2023 · 7bcd2fd · 7bcd2fd
1 parent d951858
commit 7bcd2fd
Show file tree

Hide file tree

Showing 8 changed files with 448 additions and 4,352 deletions.
diff --git a/examples/multimodal_image_search_clip.ipynb b/examples/multimodal_image_search_clip.ipynb
diff --git a/examples/sql-example.ipynb b/examples/sql-example.ipynb
@@ -62,6 +62,16 @@
     "The initial step in any `superduperdb` workflow is to connect to your datastore. To connect to a different datastore, simply add a different `URI`, for example, `postgres://...`."
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "587bffd3-7202-4e12-a6de-435d4b16892d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!rm .superduperdb/test.ddb"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -78,6 +88,16 @@
     "db = superduper('duckdb://.superduperdb/test.ddb')"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0a5252a3-91a7-4789-85b7-0b8279c228a3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "db.show('vector_index')"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "b8794451",
@@ -108,7 +128,7 @@
     "!mkdir -p data/coco\n",
     "\n",
     "# Move the 'images_small' directory to 'data/coco/images'\n",
-    "!mv images_small data/coco/images"
+    "!mv images_tiny data/coco/images"
    ]
   },
   {
@@ -224,7 +244,7 @@
    "source": [
     "## Build SuperDuperDB `Model` Instances\n",
     "\n",
-    "This use-case utilizes the `superduperdb.ext.torch` extension. Both models used output `torch` tensors, which are encoded with `tensor`:"
+    "This use-case utilizes the `superduperdb.ext.torch` extension. Both models use `torch` tensors in their output, which are encoded with `tensor`:"
    ]
   },
   {

diff --git a/examples/transfer_learning.ipynb b/examples/transfer_learning.ipynb
@@ -4,7 +4,10 @@
    "cell_type": "markdown",
    "id": "fe6fd0ab0e1ad844",
    "metadata": {
-    "collapsed": false
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
    },
    "source": [
     "# Transfer Learning with Sentence Transformers and Scikit-Learn"
@@ -14,7 +17,10 @@
    "cell_type": "markdown",
    "id": "8dcde44d942793ff",
    "metadata": {
-    "collapsed": false
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
    },
    "source": [
     "## Introduction\n",
@@ -26,7 +32,10 @@
    "cell_type": "markdown",
    "id": "1809feca8a8dca5a",
    "metadata": {
-    "collapsed": false
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
    },
    "source": [
     "## Prerequisites\n",
@@ -41,7 +50,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "!pip install superduperdb\n",
+    "# !pip install superduperdb\n",
     "!pip install ipython numpy datasets sentence-transformers"
    ]
   },
@@ -57,7 +66,10 @@
    "cell_type": "markdown",
    "id": "5379007991707d17",
    "metadata": {
-    "collapsed": false
+    "collapsed": false,
+    "jupyter": {
+     "outputs_hidden": false
+    }
    },
    "source": [
     "First, we need to establish a connection to a MongoDB datastore via SuperDuperDB. You can configure the `MongoDB_URI` based on your specific setup. \n",
@@ -84,7 +96,7 @@
     "\n",
     "# SuperDuperDB, now handles your MongoDB database\n",
     "# It just super dupers your database\n",
-    "db = superduper(mongodb_uri)\n",
+    "db = superduper(mongodb_uri, artifact_store='filesystem://./data/')\n",
     "\n",
     "# Reference a collection called transfer\n",
     "collection = Collection('transfer')"
@@ -169,10 +181,21 @@
     "    X='text',\n",
     "    db=db,\n",
     "    select=collection.find(),\n",
-    "    listen=True\n",
+    "    listen=True,\n",
+    "    show_progress_bar=True,\n",
     ")"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "aed95196-4454-4d88-beb8-dc6dddfcbe33",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "db.execute(collection.find_one())"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "68fefc17",
@@ -201,10 +224,10 @@
     "\n",
     "# Train the model on 'text' data with corresponding labels\n",
     "model.fit(\n",
-    "    X='text',\n",
+    "    X='_outputs.text.all-MiniLM-L6-v2.0',\n",
     "    y='label',\n",
     "    db=db,\n",
-    "    select=collection.find().featurize({'text': 'all-MiniLM-L6-v2'}),\n",
+    "    select=collection.find(),\n",
     ")\n"
    ]
   },
@@ -227,9 +250,9 @@
    "source": [
     "# Make predictions on 'text' data with the trained SuperDuperDB model\n",
     "model.predict(\n",
-    "    X='text',\n",
+    "    X='_outputs.text.all-MiniLM-L6-v2.0',\n",
     "    db=db,\n",
-    "    select=collection.find().featurize({'text': 'all-MiniLM-L6-v2'}),\n",
+    "    select=collection.find(),\n",
     "    listen=True,\n",
     ")"
    ]
@@ -258,16 +281,8 @@
     "print(r['text'][:100])\n",
     "\n",
     "# Print the prediction made by the SVC model stored in '_outputs'\n",
-    "print(r['_outputs']['text']['svc'])"
+    "print(r['_outputs']['text']['svc']['0'])"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "2174a0a3-c32e-4481-a301-370b569ba30c",
-   "metadata": {},
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {
@@ -286,7 +301,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.5"
+   "version": "3.11.6"
   }
  },
  "nbformat": 4,