Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix sample code and add claude v3 code generation #220

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Fix sample code and add claude v3 code generation #220

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
7406bcc
Fix sample code and add claude v3 code generation
mnmcho Apr 2, 2024
adfd3d1
fix formatting on claude v3 to match other cells
mnmcho Apr 25, 2024
2a8136a
add note about messages API for Claude 3
mnmcho Apr 25, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
201 changes: 166 additions & 35 deletions 01_Text_generation/01_code_generation_w_bedrock.ipynb
View file
Open in desktop
Expand Up @@ -100,7 +100,9 @@
"cell_type": "code",
"execution_count": null,
"id": "89a0ad24",
"metadata": {},
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"# create sales.csv file\n",
Expand Down Expand Up @@ -178,7 +180,7 @@
"- The date with the highest revenue\n",
"- Visualize monthly sales using a bar chart\n",
"\n",
"Ensure the code is syntactically correct, bug-free, optimized, not span multiple lines unnessarily, and prefer to use standard libraries. Return only python code without any surrounding text, explanation or context.\n",
"Ensure the code is syntactically correct, bug-free, optimized, not span multiple lines unnessarily, and prefer to use standard libraries that comes with Python. Return only python code without any surrounding text, explanation or context.\n",
"\n",
"Assistant:\n",
"\"\"\""
Expand All @@ -189,7 +191,7 @@
"id": "cc9784e5-5e9d-472d-8ef1-34108ee4968b",
"metadata": {},
"source": [
"Let's start by using the Anthropic Claude V2 model."
"Let's start by using the **Anthropic Claude V2** model."
]
},
{
Expand Down Expand Up @@ -224,7 +226,9 @@
"cell_type": "code",
"execution_count": null,
"id": "016a118a",
"metadata": {},
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"modelId = 'anthropic.claude-v2' # change this to use a different version from the model provider\n",
Expand All @@ -234,7 +238,9 @@
"response = boto3_bedrock.invoke_model(body=body, modelId=modelId, accept=accept, contentType=contentType)\n",
"response_body = json.loads(response.get('body').read())\n",
"\n",
"response_body.get('completion')"
"response_text = response_body.get('completion')\n",
"\n",
"print(response_text)"
]
},
{
Expand All @@ -249,52 +255,177 @@
"cell_type": "code",
"execution_count": null,
"id": "77d9b428",
"metadata": {},
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"# Sample Generated Python Code ( Generated with Amazon Bedrock in previous step)\n",
"# Sample Generated Python Code (Generated with Amazon Bedrock in previous step)\n",
"\n",
"import csv\n",
"from collections import defaultdict\n",
"import matplotlib.pyplot as plt\n",
"\n",
"revenue = 0\n",
"monthly_revenue = defaultdict(int)\n",
"product_revenue = defaultdict(int)\n",
"max_revenue = 0\n",
"max_revenue_date = ''\n",
"max_revenue_product = ''\n",
"max_revenue_product = None\n",
"max_revenue_date = None\n",
"monthly_sales = defaultdict(int)\n",
"\n",
"with open('sales.csv') as f:\n",
" reader = csv.reader(f)\n",
" next(reader)\n",
" next(reader) # skip header\n",
" for row in reader:\n",
" date = row[0]\n",
" product = row[1]\n",
" price = float(row[2])\n",
" units = int(row[3])\n",
"\n",
" revenue += price * units\n",
" product_revenue[product] += price * units\n",
" monthly_revenue[date[:7]] += price * units\n",
"\n",
" if revenue > max_revenue:\n",
" max_revenue = revenue\n",
" max_revenue_date = date\n",
" date, product, price, units = row\n",
" revenue += float(price) * int(units)\n",
" product_revenue = float(price) * int(units)\n",
" if product_revenue > max_revenue:\n",
" max_revenue = product_revenue\n",
" max_revenue_product = product\n",
" max_revenue_date = date\n",
" month = date.split('-')[1]\n",
" monthly_sales[month] += product_revenue\n",
" \n",
"print(f'Total revenue: {revenue}') \n",
"print(f'Product with highest revenue: {max_revenue_product}')\n",
"print(f'Date with highest revenue: {max_revenue_date}')\n",
"\n",
"months = list(monthly_revenue.keys())\n",
"values = list(monthly_revenue.values())\n",
"\n",
"plt.bar(months, values)\n",
"plt.bar(monthly_sales.keys(), monthly_sales.values())\n",
"plt.xlabel('Month')\n",
"plt.ylabel('Revenue')\n",
"plt.title('Monthly Revenue')\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "a9464363",
"metadata": {},
"source": [
"Now, let's try **Anthropic Claude V3** model. We will use Claude 3 Sonnet here.\n",
"\n",
"For Claude v3, [Messages API](https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages.html) must be used instead of [Text Completions API](https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-text-completion.html) that are used in the above Claude v2 example. Claude models are trained to operate on alternating user and assistant conversationl turns, so each input message must be an object with a role and content. This message object can be a single user-role message, or it can be multiple user and assistant messages.\n",
"\n",
"Messages API is designed to be backward compatible, so it works with Claude 3 models (Haiku, Sonnet, Opus) as well as Claude 2 models (v2, v2.1) and Claude Instant v1.2.\n",
"\n",
"This is why we define the `messages` variable and include `messages` parameter in the body for the following cell."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "4d9f2582-3275-4609-86f7-4c660e9c0303",
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"messages = [{\"role\": \"user\", \"content\": prompt_data}]\n",
"\n",
"body = json.dumps({\n",
" \"anthropic_version\": \"\",\n",
" \"max_tokens\": 4096,\n",
" \"messages\": messages,\n",
" \"temperature\": 0.5,\n",
" \"top_k\": 250,\n",
" \"top_p\": 0.5,\n",
" \"stop_sequences\": [\"\\n\\nHuman:\"]\n",
" })"
]
},
{
"cell_type": "markdown",
"id": "505708ec",
"metadata": {},
"source": [
"#### Invoke the Anthropic Claude v3 model"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "4337ec93",
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"modelId = 'anthropic.claude-3-sonnet-20240229-v1:0' # change this to use a different version from the model provider\n",
"\n",
"response = boto3_bedrock.invoke_model(body=body, modelId=modelId)\n",
"response_body = json.loads(response.get('body').read())\n",
"\n",
"response_text = response_body.get('content')[0]['text']\n",
"\n",
"print(response_text)"
]
},
{
"cell_type": "markdown",
"id": "0885f663",
"metadata": {},
"source": [
"#### (Optional) Execute the Bedrock generated code for validation. Go to text editor to copy the generated code as printed output can be trucncated. Replace the code in below cell."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "465755c3",
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"# Sample Generated Python Code (Generated with Amazon Bedrock in previous step)\n",
"\n",
"import csv\n",
"from collections import defaultdict\n",
"import matplotlib.pyplot as plt\n",
"\n",
"# Read data from CSV file\n",
"data = []\n",
"with open('sales.csv', 'r') as file:\n",
" reader = csv.DictReader(file)\n",
" for row in reader:\n",
" data.append(row)\n",
"\n",
"# Calculate total revenue\n",
"total_revenue = sum(float(row['price']) * float(row['units_sold']) for row in data)\n",
"\n",
"# Find product with highest revenue\n",
"product_revenue = defaultdict(float)\n",
"for row in data:\n",
" product_revenue[row['product_id']] += float(row['price']) * float(row['units_sold'])\n",
"highest_revenue_product = max(product_revenue, key=product_revenue.get)\n",
"\n",
"# Find date with highest revenue\n",
"date_revenue = defaultdict(float)\n",
"for row in data:\n",
" date_revenue[row['date']] += float(row['price']) * float(row['units_sold'])\n",
"highest_revenue_date = max(date_revenue, key=date_revenue.get)\n",
"\n",
"# Visualize monthly sales\n",
"monthly_sales = defaultdict(float)\n",
"for row in data:\n",
" year, month, _ = row['date'].split('-')\n",
" monthly_sales[f\"{year}-{month}\"] += float(row['units_sold'])\n",
"\n",
"months = sorted(monthly_sales.keys())\n",
"units_sold = [monthly_sales[month] for month in months]\n",
"\n",
"plt.figure(figsize=(10, 6))\n",
"plt.bar(months, units_sold)\n",
"plt.xlabel('Month')\n",
"plt.ylabel('Units Sold')\n",
"plt.title('Monthly Sales')\n",
"plt.xticks(rotation=90)\n",
"plt.tight_layout()\n",
"plt.show()\n",
"\n",
"print('Total Revenue:', revenue)\n",
"print('Product with max revenue:', max_revenue_product)\n",
"print('Date with max revenue:', max_revenue_date)"
"print(f\"Total revenue: ${total_revenue:.2f}\")\n",
"print(f\"Product with highest revenue: {highest_revenue_product}\")\n",
"print(f\"Date with highest revenue: {highest_revenue_date}\")"
]
},
{
Expand Down Expand Up @@ -428,7 +559,7 @@
{
"cell_type": "code",
"execution_count": null,
"id": "fe7ae02d-4b4e-470a-bfbe-9cbdd30b8db6",
"id": "3289c9d3-ea79-4739-aa6b-4ad2390951b1",
"metadata": {},
"outputs": [],
"source": []
Expand Down Expand Up @@ -1042,9 +1173,9 @@
],
"instance_type": "ml.t3.medium",
"kernelspec": {
"display_name": "Python 3 (Data Science 3.0)",
"display_name": "conda_python3",
"language": "python",
"name": "python3__SAGEMAKER_INTERNAL__arn:aws:sagemaker:us-east-1:081325390199:image/sagemaker-data-science-310-v1"
"name": "conda_python3"
},
"language_info": {
"codemirror_mode": {
Expand All @@ -1056,7 +1187,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.6"
"version": "3.10.13"
},
"vscode": {
"interpreter": {
Expand Down