Add auto structure improvement

app.py

@@ -286,6 +286,13 @@ Included words: {included_phrases}
 Excluded words: {excluded_phrases}"""
 
 
+improve_structure_prompt = """You are given a structure for a product description. 
+Reformat the structure so that it is easier to read and understand for human and AI. 
+Return the reformatted structure only. Do not add any preceding or trailing characters.
+
+<structure>/n{structure}</structure>"""
+
+
 import base64
 import requests
 
@@ -433,7 +440,7 @@ Do not add any preceding or trailing text to the output.
 Product description: 
 {text}
 
-Guildlines: 
+Guidelines: 
 {guidance_prompt}""")
     ]
 
@@ -500,6 +507,15 @@ Text to rewrite:
     return ""
 
 
+def improve_structure(chat: ChatOpenAI, structure: str):
+    messages = [
+        HumanMessage(content=improve_structure_prompt.format(structure=structure)),]
+    response = chat.invoke(messages)
+    print("Original structure: ", structure)
+    print("Improved structure: ", response.content)
+    return response.content
+
+
 def generate(*data):
     global visible
     print("visible", visible)
@@ -564,6 +580,8 @@ def generate(*data):
             if len(copy.strip()) > 0 and len(structure.strip()) > 0:    
                 print('------------')
                 print("Using both copy and structure")
+                print("Improving structure")
+                structure = improve_structure(chat=chat, structure=structure)
                 messages = [
                     SystemMessage(content=f"""You are a helpful assistant that writes about products for ecommerce websites. Make sure to write in {languages[i]} language."""),
                     HumanMessage(content=struct_copy_prompt.format(nversions=nversions, min_length=min_length, max_length=max_length, key_features=key_features, structure=structure, copy=copy, included_phrases=included_phrases, excluded_phrases=excluded_phrases)),]
@@ -748,7 +766,7 @@ def authf(username, password):
 if __name__ == '__main__':
     # demo.launch(server_name="0.0.0.0", auth=authf)
     with open("formatted_prompts.txt", "w") as f:
-        f.writelines('"write_struct_copy": "' + str2escaped(struct_copy_prompt).strip() + '",')
-        f.writelines('"write_copy": "' + str2escaped(copy_prompt).strip() + '",')
-        f.writelines('"write_struct": "' + str2escaped(struct_prompt).strip() + '",')
+        f.writelines('"write_struct_copy": "' + str2escaped(struct_copy_prompt).strip() + '",\n')
+        f.writelines('"write_copy": "' + str2escaped(copy_prompt).strip() + '",\n')
+        f.writelines('"write_struct": "' + str2escaped(struct_prompt).strip() + '",\n')
     demo.launch()

str2escaped.py

@@ -193,15 +193,26 @@ Excluded words: {excluded_words}"""
 
 test_prompt = "Generate {nversions} versions of the product description for a product with the following information.\nWrite in a way that target the customer.\nFor each version, write at least {min_length} words and at most {max_length} words.\nMake sure that the structure of each output follows the reference structure. Do not add any additional sentences or structures that are not in the reference structure.\nMake sure to use the tone of voice, rythm, cadence and style of the reference copy for each output.\nUse markdown format for each output.\nDo not include any part of the reference structure in the output.\nThe structure of the output should follow the reference structure.\nDo not use the structure of the reference copy in the output.\nDo not use any of the excluded words in the output.\nInclude all included words in the output.\nDo not hallucinate any information. Do not hallucinate information about size and weight. Write about size and weight only if it is available in the list of features.\nUse creative language in each output, do not use the common ways of starting product descriptions.\nAvoid common phrases and cliches such as \"Step into something\", \"Elevate something\", \"Discover something\", \"Unleash something\", \"Embrace something\", and similar phrases.\nFor each version, try to write in different style.\nPut each item in the list on a separate line. Use '-' character to start each item. \nRate the quality of each version based on the following criteria:\n    - how well it follows the reference structure.\n    - how well it follows the tone of voice, rythm, cadence and style of the reference copy.\n    - how faithful it describes the product features.\n    - how well it avoid the excluded words.\n    - how well it includes the included words.\n    - how creative the language is.\nThe score should be a number between 0 and 10 with 10 being the best quality.\nReturn the result in the following JSON format:\n{{\n    \"versions\": [\n        {{\n            \"id\": 1,\n            \"content\": The first product description,\n            \"explanation\": A less than 20 word explanation of the score of the first product description,\n            \"score\": The score of the first product description\n        }},\n        {{\n            \"id\": 2,\n            \"content\": The second product description,\n            \"explanation\": A less than 20 word explanation of the score of the first product description,\n            \"score\": The score of the second product description\n        }},\n        ...\n    ],\n    \"best_version\": {{\n        \"explanation\": Explanation for why this version is the best,\n        \"id\": The id of the best version\n    }}\n}}\nMake sure that the output is in JSON format, no extra text should be included in the output.\n\nProduct information:\nKey features: {key_features}\n\nReference structure: {structure}\n\nReference copy: {copy}\n\nIncluded words: {included_phrases}\n\nExcluded words: {excluded_phrases}"
 
+detect_feature = """Describe the features of the {garment_type} in the photo in less than 150 words.
+Return the result in in the following JSON format without any preceding or trailing text: 
+
+{{
+    "features": [list of comma separated features], 
+    "intended_use": [list of comma separated intended uses], 
+    "alt_text": alt text for the image,
+    "category": the category of the garment, for example 'dress', 'shirt', 'pants', 'shoes', etc
+}}"""
+
 import codecs
 if __name__ == "__main__":
-    print()
-    print('"write_struct_copy": "' + str2escaped(both).strip() + '",')
-    print()
-    print('"write_copy": "' + str2escaped(copy).strip() + '",')
-    print()
-    print('"write_struct": "' + str2escaped(struct).strip() + '",')
-    print()
+    print(str2escaped(detect_feature))
+    # print()
+    # print('"write_struct_copy": "' + str2escaped(both).strip() + '",')
+    # print()
+    # print('"write_copy": "' + str2escaped(copy).strip() + '",')
+    # print()
+    # print('"write_struct": "' + str2escaped(struct).strip() + '",')
+    # print()
     # print('test\n', str2escaped(test))