style: remove autoreload, format

xtreamsrl · Feb 12, 2025 · fc4e61e · fc4e61e
1 parent df69e1f
commit fc4e61e
Showing 1 changed file with 16 additions and 27 deletions.
diff --git a/notebooks/02-domain-expert.ipynb b/notebooks/02-domain-expert.ipynb
@@ -33,7 +33,7 @@
    "id": "3",
    "metadata": {},
    "source": [
-    "Now that we have chosen our SME, what could we ask him/her? We could ask him to judge some answers our system gives to a list of questions. We can leverage the power of LLM to invent questions for use, this could be useful to bootstrap our evaluation pipeline. "
+    "Now that we have chosen our SME, what could we ask him/her? We could ask him to judge some answers our system gives to a list of questions. We can leverage the power of LLM to invent questions for use, this could be useful to bootstrap our evaluation pipeline."
    ]
   },
   {
@@ -42,17 +42,6 @@
    "id": "4",
    "metadata": {},
    "outputs": [],
-   "source": [
-    "%load_ext autoreload\n",
-    "%autoreload 2"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "5",
-   "metadata": {},
-   "outputs": [],
    "source": [
     "from pydantic import BaseModel\n",
     "import polars as pl"
@@ -61,7 +50,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "6",
+   "id": "5",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -72,7 +61,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "7",
+   "id": "6",
    "metadata": {},
    "source": [
     "To instruct the LLM to generate valuable questions to use as an evaluation dataset we could think about different characteristics that our product must have, in particular: \n",
@@ -86,7 +75,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "8",
+   "id": "7",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -106,7 +95,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "9",
+   "id": "8",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -121,13 +110,15 @@
     "\"\"\"\n",
     "\n",
     "MAX_ROWS = 20\n",
-    "EVAL_CONSTRUCTION_PROMPT = \"\"\"Generate an evaluation dataset with no more than {n_rows} rows\"\"\""
+    "EVAL_CONSTRUCTION_PROMPT = (\n",
+    "    \"\"\"Generate an evaluation dataset with no more than {n_rows} rows\"\"\"\n",
+    ")"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "10",
+   "id": "9",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -149,7 +140,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "11",
+   "id": "10",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -162,7 +153,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "12",
+   "id": "11",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -181,16 +172,14 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "13",
+   "id": "12",
    "metadata": {},
    "outputs": [],
    "source": [
     "import openai\n",
     "\n",
-    "\n",
     "client = openai.OpenAI()\n",
     "\n",
-    "\n",
     "chat_completion = client.beta.chat.completions.parse(\n",
     "    model=\"gpt-4o\",\n",
     "    messages=[\n",
@@ -208,7 +197,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "14",
+   "id": "13",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -217,7 +206,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "15",
+   "id": "14",
    "metadata": {},
    "source": [
     "Now, we have a list of questions to pose to our AI and ask our evaluation expert to evaluate it. Note that, the SME could be involved also both for giving you features, scenario and personas or to add particular questions to the generated dataset. "
@@ -226,7 +215,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "16",
+   "id": "15",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -237,7 +226,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "17",
+   "id": "16",
    "metadata": {},
    "outputs": [],
    "source": [