feat(#1132): Introduce async version of our API

docs/guides/monitoring.ipynb

@@ -216,6 +216,120 @@
     "dataset.map(make_prediction)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "6987c362-61b3-4682-aa7b-693bed30d3ae",
+   "metadata": {},
+   "source": [
+    "## Using `rb.log` in background mode\n",
+    "\n",
+    "You can monitor your own models without adding a response delay by using the `background` param in rb.log\n",
+    "\n",
+    "Let's see an example using [BentoML](https://www.bentoml.com/) with a spaCy NER pipeline:\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5a6e24b1-b665-4e44-be52-62e024927643",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import spacy\n",
+    "\n",
+    "nlp = spacy.load(\"en_core_web_sm\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "45740aee-e022-452d-8f77-fb8872eb7d11",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%%writefile spacy_model.py\n",
+    "\n",
+    "\n",
+    "from bentoml import BentoService, api, artifacts, env\n",
+    "from bentoml.adapters import JsonInput\n",
+    "from bentoml.frameworks.spacy import SpacyModelArtifact\n",
+    "\n",
+    "import rubrix as rb\n",
+    "\n",
+    "\n",
+    "@env(infer_pip_packages=True)\n",
+    "@artifacts([SpacyModelArtifact(\"nlp\")])\n",
+    "class SpacyNERService(BentoService):\n",
+    "\n",
+    "    @api(input=JsonInput(), batch=True)\n",
+    "    def predict(self, parsed_json_list):\n",
+    "        result, rb_records = ([], [])\n",
+    "        for index, parsed_json in enumerate(parsed_json_list):\n",
+    "            doc = self.artifacts.nlp(parsed_json[\"text\"])\n",
+    "            prediction = [{\"entity\": ent.text, \"label\": ent.label_} for ent in doc.ents]\n",
+    "            rb_records.append(\n",
+    "                rb.TokenClassificationRecord(\n",
+    "                    text=doc.text,\n",
+    "                    tokens=[t.text for t in doc],\n",
+    "                    prediction=[\n",
+    "                        (ent.label_, ent.start_char, ent.end_char) for ent in doc.ents\n",
+    "                    ],\n",
+    "                )\n",
+    "            )\n",
+    "            result.append(prediction)\n",
+    "        \n",
+    "        rb.log(\n",
+    "            name=\"monitor-for-spacy-ner\",\n",
+    "            records=rb_records,\n",
+    "            tags={\"framework\": \"bentoml\"},\n",
+    "            background=True, \n",
+    "            verbose=False\n",
+    "        ) # By using the background=True, the model latency won't be affected\n",
+    "        \n",
+    "        return result\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "126b826f-c708-4b1d-898d-960058a7635d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from spacy_model import SpacyNERService\n",
+    "\n",
+    "svc = SpacyNERService()\n",
+    "svc.pack('nlp', nlp)\n",
+    "\n",
+    "saved_path = svc.save()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b5fadd39-87ed-4b7c-9f62-7cfdba0e2813",
+   "metadata": {},
+   "source": [
+    "You can predict some data without serving the model. Just launch following command:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "69736fa1-e263-4086-9a53-a0be399b22a1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!bentoml run SpacyNERService:latest predict --input \"{\\\"text\\\":\\\"I am driving BMW\\\"}\""
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "8a1fe7dd-3ef0-42c6-bb2b-25a52a95c582",
+   "metadata": {},
+   "source": [
+    "If you're running Rubrix in local, go to http://localhost:6900/datasets/rubrix/monitor-for-spacy-ner and see that the new dataset `monitor-for-spacy-ner` contains your data"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "c71b49ea-7384-423d-b2d8-ac6280b7a200",
@@ -229,7 +343,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "Python 3 (ipykernel)",
    "language": "python",
    "name": "python3"
   },
@@ -243,9 +357,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.5"
+   "version": "3.8.12"
   }
  },
  "nbformat": 4,
  "nbformat_minor": 5
-}
+}

src/rubrix/__init__.py

@@ -36,6 +36,7 @@
         init,
         load,
         log,
+        log_async,
         set_workspace,
     )
     from rubrix.client.datasets import (
@@ -62,6 +63,7 @@
         "init",
         "load",
         "log",
+        "log_async",
         "set_workspace",
     ],
     "client.models": [