🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org
Python agentai-societiesartificial-intelligencecommunicative-aicooperative-aideep-learninglarge-language-modelsmulti-agent-systemsnatural-language-processing
usecase: Add pptx_toolkit usecase app (#2562)
Sarthak Bhardwaj9f7f252344f4
diff --git a/examples/usecases/mistral_OCR/README.md b/examples/usecases/mistral_OCR/README.mdindex d8a899790..3f1d71abf 100644--- a/examples/usecases/mistral_OCR/README.md+++ b/examples/usecases/mistral_OCR/README.md@@ -84,5 +84,5 @@ requirements.txt # Python dependenciesstreamlitpillowmistralai-camel-ai==0.2.61+camel-ai[all]==0.2.61```diff --git a/examples/usecases/mistral_OCR/requirements.txt b/examples/usecases/mistral_OCR/requirements.txtindex 52815067d..41e9bc444 100644--- a/examples/usecases/mistral_OCR/requirements.txt+++ b/examples/usecases/mistral_OCR/requirements.txt@@ -1,4 +1,4 @@-camel-ai==0.2.61+camel-ai[all]==0.2.61streamlitpillowmistralaidiff --git a/examples/usecases/pptx_toolkit_usecase/README.md b/examples/usecases/pptx_toolkit_usecase/README.mdnew file mode 100644index 000000000..e4bedb65e--- /dev/null+++ b/examples/usecases/pptx_toolkit_usecase/README.md@@ -0,0 +1,100 @@+# 📑 AI-Powered PPTX Generator (CAMEL-AI)++Generate beautiful, professional PowerPoint presentations (PPTX) on any topic in seconds!+Built with [CAMEL-AI](https://github.com/camel-ai/camel), OpenAI, and Streamlit.++---++## 🚀 Features++- 🔥 Generate engaging, multi-slide PPTX files for any topic—instantly!+- 💡 Uses OpenAI to create bullet, step-by-step (pentagon/chevron), and table slides.+- 🖼️ Optional support for auto-inserting images using [Pexels](https://www.pexels.com/api/).+- 📦 All output is 100% local and ready for download.++---++## 🛠️ Installation++```bash+# 1. Create a fresh virtual environment (recommended)+python3 -m venv .venv+source .venv/bin/activate # On Windows: .venv\Scripts\activate++# 2. Install the specific CAMEL-AI version (and requirements)+pip install camel-ai[all]==0.2.62 streamlit openai++# 3. (Optional) Get a [Pexels API key](https://www.pexels.com/api/) if you want slides with images+```++---++## 📝 Usage++1. **Get your OpenAI API key:**+ [Create one here](https://platform.openai.com/account/api-keys).++2. **(Optional) Get your Pexels API key:**+ [Request it here](https://www.pexels.com/api/) if you want to auto-insert images.++3. **Clone this repo or copy the app file:**+ ```bash+ git clone https://github.com/camel-ai/camel+ cd examples/usecases/pptx_toolkit_usecase+ # Or just copy the .py script to your working directory+ ```++4. **Run the Streamlit app:**+ ```bash+ streamlit run app_pptx.py+ ```++5. **Open the Streamlit UI in your browser** (default: http://localhost:8501).++6. **Enter your API keys, type your topic, set slide count, and hit "Generate Presentation"!**++---++## 🏗️ How it Works++- You provide a topic and desired slide count.+- The app prompts OpenAI GPT-4o or GPT-4 (you can change the model in code) to generate slide content in a format CAMEL-AI's `PPTXToolkit` can use.+- At least one step-by-step slide and one table slide are always included; at least two slides will contain images if you provide a Pexels API key.+- The toolkit generates a `.pptx` file named after your topic, ready for download.++---++## ⚙️ Example Prompt Structure++<details>+<summary>See Example</summary>++```json+[+ {"title": "AI Agents", "subtitle": "Exploring the world of artificial intelligence agents"},+ {"heading": "Types of AI Agents", "bullet_points": ["Intelligent Virtual Agents", "Autonomous Agents", "Collaborative Agents"], "img_keywords": "AI, technology"},+ {"heading": "Creating an AI Agent", "bullet_points": [">> Step 1: Define the goal", ">> Step 2: Choose algorithms", ">> Step 3: Implement and test"], "img_keywords": "workflow, robotics"},+ {"heading": "Comparison of AI Agents", "table": {"headers": ["Type", "Capabilities", "Examples"], "rows": [["Virtual", "Conversational AI", "Siri"], ["Autonomous", "Self-learning", "Robots"]]}, "img_keywords": "comparison chart, table"}+]+```+</details>++---++## 📝 Credits++- Built with [CAMEL-AI](https://github.com/camel-ai/camel) (PPTXToolkit module)+- Powered by [OpenAI](https://platform.openai.com/docs/api-reference) (for content generation)+- Images by [Pexels](https://www.pexels.com/api/) (optional)+- UI by [Streamlit](https://streamlit.io/)++---++## 🐪 License++Apache 2.0+(c) 2023-2024 [CAMEL-AI.org](https://camel-ai.org)++---++*Made with ❤️ by the CAMEL-AI community.*diff --git a/examples/usecases/pptx_toolkit_usecase/app_pptx.py b/examples/usecases/pptx_toolkit_usecase/app_pptx.pynew file mode 100644index 000000000..859985eaa--- /dev/null+++ b/examples/usecases/pptx_toolkit_usecase/app_pptx.py@@ -0,0 +1,205 @@+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========+# Licensed under the Apache License, Version 2.0 (the "License");+# you may not use this file except in compliance with the License.+# You may obtain a copy of the License at+#+# http://www.apache.org/licenses/LICENSE-2.0+#+# Unless required by applicable law or agreed to in writing, software+# distributed under the License is distributed on an "AS IS" BASIS,+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.+# See the License for the specific language governing permissions and+# limitations under the License.+# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========+import os+import streamlit as st+import json+import base64+from typing import List, Optional, Union+from pydantic import BaseModel, Field+++from camel.toolkits.pptx_toolkit import PPTXToolkit+from camel.models import ModelFactory+from camel.types import ModelPlatformType, ModelType+from camel.agents import ChatAgent++# --- Pydantic Models for Structured Output ---+class TitleSlide(BaseModel):+ """Title slide model"""+ title: str = Field(description="Main title of the presentation")+ subtitle: str = Field(description="Subtitle or description")++class TableData(BaseModel):+ """Table data model"""+ headers: List[str] = Field(description="Table column headers")+ rows: List[List[str]] = Field(description="Table rows data")++class BulletSlide(BaseModel):+ """Bullet point slide model"""+ heading: str = Field(description="Slide heading")+ bullet_points: List[str] = Field(description="List of bullet points, use >> prefix for step-by-step slides")+ img_keywords: Optional[str] = Field(default="", description="Keywords for image search (not URLs)")++class TableSlide(BaseModel):+ """Table slide model"""+ heading: str = Field(description="Slide heading")+ table: TableData = Field(description="Table data with headers and rows")+ img_keywords: Optional[str] = Field(default="", description="Keywords for image search (not URLs)")++class PresentationSlides(BaseModel):+ """Complete presentation model"""+ title_slide: TitleSlide = Field(description="First slide must be a title slide")+ content_slides: List[Union[BulletSlide, TableSlide]] = Field(+ description="Content slides including bullet slides and table slides"+ )+++# Load camel logo as base64+def get_base64_img(path, w=32):+ with open(path, "rb") as img_f:+ b64 = base64.b64encode(img_f.read()).decode()+ return f"<img src='data:image/png;base64,{b64}' width='{w}' style='vertical-align:middle;margin-bottom:4px;'>"++camel_logo = get_base64_img("assets/CAMEL_logo.jpg", 40)++# --- Page config ---+st.markdown(+ f"""+ <div style='display:flex; align-items:center; gap:12px;'>+ <span style="font-size:2.2rem;font-weight:700;">PPTX Generator using pptx_toolkit with</span>+ {camel_logo}+ </div>+ """,+ unsafe_allow_html=True+)++# --- Sidebar Config ---+st.sidebar.header("⚙️ Configuration")+openai_key = st.sidebar.text_input(+ "OpenAI API Key (for ChatAgent/OpenAI backend)", type="password",+ help="Get yours at https://platform.openai.com/account/api-keys"+)+pexels_key = st.sidebar.text_input(+ "Pexels API Key", type="password", help="(Optional) Needed if you want images fetched from Pexels"+)++# --- User Input ---+st.header("1. Enter Your Presentation Topic")+topic = st.text_input("Topic (e.g., ‘Blockchain for Beginners’)")+slide_count = st.slider("Number of slides (excluding title slide)", 3, 10, 5)++# --- Construct the JSON‐generation instructions ---+def pptx_prompt(topic: str, slide_count: int) -> str:+ r"""Create a prompt for structured output generation"""+ return f"""+Create a presentation about "{topic}" with exactly {slide_count + 1} slides total (1 title slide + {slide_count} content slides).+Requirements:+1. First slide must be a title slide with appropriate title and subtitle+2. Include at least one step-by-step slide (bullet points starting with ">>")+3. Include at least one table slide with relevant data+4. At least TWO content slides must have meaningful img_keywords for visual content+5. Use Markdown formatting (**bold**, *italic*) in bullet points+6. Make content clear, concise, and engaging for the topic "{topic}"+The presentation should cover key aspects of {topic}, including practical information, processes, and relevant data.+"""++# --- Generate Slide JSON via ChatAgent ---+def generate_pptx_json_with_agent(topic: str, slide_count: int, api_key: str):+ # 1. Build the full instruction prompt+ full_prompt = pptx_prompt(topic, slide_count)++ # 2. Instantiate a ChatAgent pointing at OpenAI GPT-4o+ os.environ["OPENAI_API_KEY"] = api_key+ agent = ChatAgent(+ system_message="You are an AI Assistant that strictly follows instructions to produce only valid JSON for slides.",+ message_window_size=5,+ model=ModelFactory.create(+ model_platform=ModelPlatformType.OPENAI,+ model_type=ModelType.GPT_4_1,+ model_config_dict={"temperature": 0.0},+ )+ )++ # 3. Call the agent with our instruction prompt+ try:+ response = agent.step(full_prompt,response_format=PresentationSlides)++ # Parse JSON string into Pydantic object+ json_content = response.msgs[0].content+ presentation_data = PresentationSlides.model_validate_json(json_content)++ # Convert to the expected JSON format+ slides_json = []++ # Add title slide+ slides_json.append({+ "title": presentation_data.title_slide.title,+ "subtitle": presentation_data.title_slide.subtitle+ })++ # Add content slides+ for slide in presentation_data.content_slides:+ if isinstance(slide, BulletSlide):+ slides_json.append({+ "heading": slide.heading,+ "bullet_points": slide.bullet_points,+ "img_keywords": slide.img_keywords or ""+ })+ elif isinstance(slide, TableSlide):+ slides_json.append({+ "heading": slide.heading,+ "table": {+ "headers": slide.table.headers,+ "rows": slide.table.rows+ },+ "img_keywords": slide.img_keywords or ""+ })++ return slides_json, None++ except Exception as e:+ return None, f"❌ Structured output generation failed: {e}"++# --- Build & Return PPTX Bytes ---+def build_pptx(slides):+ os.environ["PEXELS_API_KEY"] = pexels_key or ""+ pptx_toolkit = PPTXToolkit(output_dir="outputs")++ out_name = f"demo_pptx_{abs(hash(json.dumps(slides)))}.pptx"+ result = pptx_toolkit.create_presentation(json.dumps(slides), out_name)+ pptx_path = os.path.join("outputs", out_name)+ if os.path.exists(pptx_path):+ with open(pptx_path, "rb") as f:+ return f.read(), out_name, result+ else:+ return None, None, result++# --- Main App Logic ---+if not openai_key:+ st.info("Please enter your OpenAI API key in the sidebar to continue.")+ st.stop()++if topic and st.button("Generate Presentation"):+ st.info("🕒 Generating slide JSON with ChatAgent (GPT-4.1)...")+ slides, error = generate_pptx_json_with_agent(topic, slide_count, openai_key)+ if error:+ st.error(error)+ st.stop()++ st.success("✅ Slide JSON created! Now building PPTX...")+ pptx_bytes, fname, pptx_result = build_pptx(slides)++ if pptx_bytes:+ st.success("🎉 Your presentation is ready!")+ st.download_button(+ label=f"Download {fname}",+ data=pptx_bytes,+ file_name=fname,+ mime="application/vnd.openxmlformats-officedocument.presentationml.presentation"+ )+ else:+ st.error(f"PPTX generation failed: {pptx_result}")++st.markdown("---")+st.caption("Made with ❤️ using CAMEL-AI, OpenAI & PPTXToolkit")\ No newline at end of filediff --git a/examples/usecases/pptx_toolkit_usecase/assets/CAMEL_logo.jpg b/examples/usecases/pptx_toolkit_usecase/assets/CAMEL_logo.jpgnew file mode 100644index 000000000..a6332e267Binary files /dev/null and b/examples/usecases/pptx_toolkit_usecase/assets/CAMEL_logo.jpg differdiff --git a/examples/usecases/pptx_toolkit_usecase/requirements.txt b/examples/usecases/pptx_toolkit_usecase/requirements.txtnew file mode 100644index 000000000..63bfbe2f9--- /dev/null+++ b/examples/usecases/pptx_toolkit_usecase/requirements.txt@@ -0,0 +1,3 @@+camel-ai[all]==0.2.61+streamlit+python-pptx\ No newline at end of filediff --git a/examples/usecases/pptx_toolkit_usecase/slides_templates/urban_monochrome.pptx b/examples/usecases/pptx_toolkit_usecase/slides_templates/urban_monochrome.pptxnew file mode 100644index 000000000..07384e652Binary files /dev/null and b/examples/usecases/pptx_toolkit_usecase/slides_templates/urban_monochrome.pptx differdiff --git a/examples/usecases/youtube_ocr/requirements.txt b/examples/usecases/youtube_ocr/requirements.txtindex e1b759b9d..c53a9809e 100644--- a/examples/usecases/youtube_ocr/requirements.txt+++ b/examples/usecases/youtube_ocr/requirements.txt@@ -3,7 +3,7 @@ opencv-python==4.9.0.80streamlit==1.45.0pytesseract==0.3.13streamlit==1.45.1-camel-ai==0.2.61+camel-ai[all]==0.2.61python-dotenv==1.1.0Pillow==11.2.1git+https://github.com/openai/whisper.git