Add image generated by AI; also bring back the manual step for JSON step, in case a network timeout occurs

app.py

@@ -1,9 +1,9 @@
-from typing import List, Tuple
-
+import base64
 import json5
 import time
 import streamlit as st
 import streamlit.runtime.scriptrunner as st_sr
+from typing import List, Tuple
 
 import llm_helper
 import pptx_helper
@@ -55,6 +55,18 @@ def get_web_search_results_wrapper(text: str) -> List[Tuple[str, str]]:
     return results
 
 
+@st.cache_data
+def get_ai_image_wrapper(text: str) -> str:
+    """
+    Fetch and cache a Base 64-encoded image by calling an external API.
+
+    :param text: The image prompt
+    :return: The Base 64-encoded image
+    """
+
+    return llm_helper.get_ai_image(text)
+
+
 def build_ui():
     """
     Display the input elements for content generation. Only covers the first step.
@@ -89,7 +101,7 @@ def build_ui():
 
     if st.session_state.clicked[0]:
         # if desc_topic_btn_submitted:
-        progress_text = 'Generating your presentation slides...give it a moment'
+        progress_text = 'Generating contents for your slides...give it a moment'
         progress_bar = st.progress(0, text=progress_text)
 
         topic_txt = topic.strip()
@@ -150,13 +162,13 @@ def process_topic_inputs(topic: str, progress_bar):
             # Apparently, "nested button click" needs to be handled differently
             # https://playground.streamlit.app/?q=triple-button
 
-            # st.button(APP_TEXT['button_labels'][1], on_click=button_clicked, args=[1])
+            st.button(APP_TEXT['button_labels'][1], on_click=button_clicked, args=[1])
 
-            # if st.session_state.clicked[1]:
-            progress_text = 'Converting...give it a moment'
-            progress_bar = st.progress(0, text=progress_text)
+            if st.session_state.clicked[1]:
+                progress_text = 'Converting...give it a moment'
+                progress_bar = st.progress(0, text=progress_text)
 
-            process_slides_contents(slides_content, progress_bar)
+                process_slides_contents(slides_content, progress_bar)
         except ValueError as ve:
             st.error(f'Unfortunately, an error occurred: {ve}! '
                      f'Please change the text, try again later, or report it, sharing your inputs.')
@@ -253,6 +265,15 @@ def show_bonus_stuff(ppt_headers: List):
     for (title, link) in search_results:
         st.markdown(f'[{title}]({link})')
 
+    st.write('')
+    st.write(APP_TEXT['image_info'])
+    image = get_ai_image_wrapper(ppt_text)
+
+    if len(image) > 0:
+        image = base64.b64decode(image)
+        st.image(image, caption=ppt_text)
+        st.info('Tip: Right-click on the image to save it.', icon="ℹ️")
+
 
 def button_clicked(button):
     """

global_config.py

@@ -17,6 +17,11 @@ class GlobalConfig:
     CLARIFAI_APP_ID_GPT = 'chat-completion'
     CLARIFAI_MODEL_ID_GPT = 'GPT-3_5-turbo'
 
+    CLARIFAI_USER_ID_SD = 'stability-ai'
+    CLARIFAI_APP_ID_SD = 'stable-diffusion-2'
+    CLARIFAI_MODEL_ID_SD = 'stable-diffusion-xl'
+    CLARIFAI_MODEL_VERSION_ID_SD = '0c919cc1edfc455dbc96207753f178d7'
+
     # LLM_MODEL_TEMPERATURE: float = 0.5
     LLM_MODEL_MIN_OUTPUT_LENGTH: int = 50
     LLM_MODEL_MAX_OUTPUT_LENGTH: int = 2000
@@ -32,15 +37,15 @@ class GlobalConfig:
     PPTX_TEMPLATE_FILES = {
         'Blank': {
             'file': 'pptx_templates/Blank.pptx',
-            'caption': 'Good for starting'
+            'caption': 'A good start'
         },
         'Ion Boardroom': {
             'file': 'pptx_templates/Ion_Boardroom.pptx',
-            'caption': 'Choose some bold colors'
+            'caption': 'Make some bold decisions'
         },
         'Urban Monochrome': {
             'file': 'pptx_templates/Urban_monochrome.pptx',
-            'caption': 'Bring in some elegance'
+            'caption': 'Marvel in a monochrome dream'
         }
     }
 

llm_helper.py

@@ -1,4 +1,7 @@
+import json
+import time
 import metaphor_python as metaphor
+import requests
 from langchain import PromptTemplate
 from langchain.llms import Clarifai
 
@@ -180,9 +183,61 @@ def get_related_websites(query: str) -> metaphor.api.SearchResponse:
     return metaphor_client.search(query, use_autoprompt=True, num_results=5)
 
 
-if __name__ == '__main__':
-    results = get_related_websites('5G AI WiFi 6')
+def get_ai_image(text: str) -> str:
+    """
+    Get a Stable Diffusion-generated image based on a given text.
 
-    for a_result in results.results:
-        print(a_result.title, a_result.url, a_result.extract)
+    :param text: The input text
+    :return: The Base 64-encoded image
+    """
+
+    url = f'''https://api.clarifai.com/v2/users/{GlobalConfig.CLARIFAI_USER_ID_SD}/apps/{GlobalConfig.CLARIFAI_APP_ID_SD}/models/{GlobalConfig.CLARIFAI_MODEL_ID_SD}/versions/{GlobalConfig.CLARIFAI_MODEL_VERSION_ID_SD}/outputs'''
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f'Key {GlobalConfig.CLARIFAI_PAT}'
+    }
+    data = {
+        "inputs": [
+            {
+                "data": {
+                    "text": {
+                        "raw": text
+                    }
+                }
+            }
+        ]
+    }
+
+    print('*** AI image generator...')
+    print(url)
+
+    start = time.time()
+    response = requests.post(
+        url=url,
+        headers=headers,
+        data=json.dumps(data)
+    )
+    stop = time.time()
+
+    print('Response:', response, response.status_code)
+    print('Image generation took', stop - start, 'seconds')
+    img_data = ''
+
+    if response.ok:
+        print('*** Clarifai SDXL request: Response OK')
+        json_data = json.loads(response.text)
+        img_data = json_data['outputs'][0]['data']['image']['base64']
+    else:
+        print('Image generation failed:', response.text)
+
+    return img_data
+
+
+if __name__ == '__main__':
+    # results = get_related_websites('5G AI WiFi 6')
+    #
+    # for a_result in results.results:
+    #     print(a_result.title, a_result.url, a_result.extract)
 
+    # get_ai_image('A talk on AI, covering pros and cons')
+    pass

requirements.txt

@@ -7,4 +7,6 @@ clarifai==9.7.4
 python-pptx
 metaphor-python
 json5~=0.9.14
-PyYAML~=6.0.1
+PyYAML~=6.0.1
+# curlify
+requests~=2.31.0

strings.json

@@ -8,10 +8,10 @@
         "Bonus Materials"
     ],
     "section_captions": [
-        "Let's start by generating some contents for your slides",
-        "Let's now convert the above generated contents into JSON",
-        "Let's now create the slides for you",
-        "Wait, there's some more good stuff for you!"
+        "Let's start by generating some contents for your slides.",
+        "Let's now convert the above generated contents into JSON.",
+        "Let's now create the slides for you.",
+        "Since you have come this far, we have unlocked some more good stuff for you!"
     ],
     "input_labels": [
         "**Describe the topic of the presentation. Avoid mentioning the count of slides.**\n*Note: the output may be short or truncated due to API limitations.*"
@@ -22,6 +22,7 @@
         "Make the slides"
     ],
     "urls_info": "Here is a list of some online resources that you can consult for further information on this topic:",
+    "image_info": "Got some more minutes? We are also trying to deliver an AI-generated art on the presentation topic, fresh off the studio, just for you!",
     "content_generation_failure_error": "Unfortunately, SlideDeck AI failed to generate any content for you! Please try again later.",
     "tos": "SlideDeck AI is an experimental prototype, and it has its limitations.\nPlease carefully review any and all AI-generated content.",
     "tos2": "By using SlideDeck AI, you agree to fair and responsible usage.\nNo liability assumed by any party."