vector-vault 5.2.1__tar.gz → 5.2.2__tar.gz

{vector_vault-5.2.1 → vector_vault-5.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vector_vault
-Version: 5.2.1
+Version: 5.2.2
 Summary: Quickly create ChatGPT RAG apps and Unleash the full potential of GenAI with Vector Vault
 Home-page: https://github.com/John-Rood/VectorVault
 Author: VectorVault.io

{vector_vault-5.2.1 → vector_vault-5.2.2}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 setup(
     name="vector_vault",
-    version="5.2.1",
+    version="5.2.2",
     packages=find_packages(),
     author="VectorVault.io",
     author_email="john@johnrood.com",

{vector_vault-5.2.1 → vector_vault-5.2.2}/vector_vault.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vector-vault
-Version: 5.2.1
+Version: 5.2.2
 Summary: Quickly create ChatGPT RAG apps and Unleash the full potential of GenAI with Vector Vault
 Home-page: https://github.com/John-Rood/VectorVault
 Author: VectorVault.io

{vector_vault-5.2.1 → vector_vault-5.2.2}/vectorvault/tools_gpt.py

@@ -3,22 +3,22 @@ import queue
 import threading
 
 '''
-    ToolsGPT is a set of tools special to large language models. 
+    ToolsGPT is a set of tools special to large language models.
     Every tool turns unstructured data into structured data.
-    
-    Structure data is required to run programs. 
+
+    Structure data is required to run programs.
     Unstructured data is the primary human output.
-    AI tools convert unstructured data into structured data, 
-    bridging the gap between humans and programs. 
+    AI tools convert unstructured data into structured data,
+    bridging the gap between humans and programs.
     Therefore, in the future, everything will be programmable.
-    
+
     ToolsGPT allows you to get the following for any input:
         1. `get_rating` - returns a rating out of 10 for any input
         2. `get_yes_no` - returns a 'yes' or a 'no' to any question
         3. `get_number` - returns an integer based on your instructions
         4. `get_match` - returns a match to a list options - (single choice)
         5. `get_multi_match` - returns one or many matches to a list of options - (multiple choice)
-        6. `get_topic` - returns the topic subject matter of any input 
+        6. `get_topic` - returns the topic subject matter of any input
 
         1. `get_rating`:
             Get a numeric rating out of 10 for anything
@@ -28,15 +28,15 @@ import threading
 
         3. `get_number`:
             Get an integer output for any input
-        
+
         4. `get_match`:
-            Get an exact match to a single item within a list 
-            -> in: (text and list of strings) 
+            Get an exact match to a single item within a list
+            -> in: (text and list of strings)
             -> out: (one exact match to an answer in list -> string type)
-        
+
         5. `get_multi_match`:
-            Get many matches to items within a list  
-            -> in: (text and list of answers) 
+            Get many matches to items within a list
+            -> in: (text and list of answers)
             -> out: (list of exact matches -> list type)
 
         6. `get_topic`:
@@ -71,7 +71,7 @@ that this number has been carefully considered given your content. My rating out
 Concept to rate out of ten for: {concept_to_rate_for} \n\n Reponse: Since I am not allowed to give any explination before or after my rating, I'd like to say
 that this number has been carefully considered given your concept. My rating out of 10 integer is: """
             prompt = prompt_template.format(concept_to_rate_for=concept_to_rate_for)
-        
+
         answer = self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
         if self.verbose == True:
             print(f"'Rating Initial Answer: {answer}")
@@ -80,18 +80,18 @@ that this number has been carefully considered given your concept. My rating out
             return int(answer) # try to return an int zero shot
         except:
             return self.retry_until_its_a_number(answer, model=model, loop_limit=loop_limit, timeout=timeout) # force the integer if zero shot fails
-        
-        
+
+
     def get_number(self, concept, content: str, model=None, loop_limit=5, timeout=None, temperature=None) -> int:
         '''
-            param: `concept` - is the idea used to generate a number with - i.e. "How many students are in the professor's class?" 
-            param: `content` - is the content to generate a number for - i.e. "I had 400 students in my class last year, but this year, I have 10% more." 
+            param: `concept` - is the idea used to generate a number with - i.e. "How many students are in the professor's class?"
+            param: `content` - is the content to generate a number for - i.e. "I had 400 students in my class last year, but this year, I have 10% more."
         '''
         response = self.retry_llm(f'{concept} \n\n{content}', model=model, loop_limit=loop_limit, timeout=timeout)
         print("Initial number extraction response:", response) if self.verbose else 0
         return response if type(response) is int else self.retry_until_its_a_number(response, model=model, loop_limit=loop_limit)
-        
-    
+
+
     def retry_until_its_a_number(self, content: str, model=None, loop_limit=5, timeout=None, temperature=None) -> int:
         '''
             param: `content` - the AI will return an integer based on the content you input
@@ -113,35 +113,35 @@ User: The following content should be a number - Content: "{content}"
         print("Initial number extraction response:", response) if self.verbose else 0
 
         if type(response) is int:
-            return response 
-        
+            return response
+
         else: # loop until its a number
             loops = 0
             answer = content
 
-            while True: 
+            while True:
                 if loops >= loop_limit:
                     break
 
                 prompt = prompt_template.format(content=answer)
                 answer = self.retry_llm(custom_prompt=prompt, model=model, loop_limit=loop_limit, timeout=timeout, temperature=temperature)
                 print(f"Loops: {loops} |  Number: {answer}") if self.verbose else 0
-                    
+
                 try:
                     answer = int(answer)
                     return answer # exit loop and return the integer
                 except:
                     loops += 1
 
-    
+
     def get_yes_no(self, text: str, question: str = None, model=None, loop_limit=20, timeout=None, temperature=None) -> str:
         '''
-            Get an exact "yes" or "no" to any question, given an input. 
+            Get an exact "yes" or "no" to any question, given an input.
             Be sure to input text to get a yes or no to, then ask the question to answer
         '''
-        answer = self.yay_or_nay(text, question, model=model, timeout=timeout) if question else self.yay_or_nay_question_in_content(text, model=model, timeout=timeout, temperature=temperature) 
+        answer = self.yay_or_nay(text, question, model=model, timeout=timeout) if question else self.yay_or_nay_question_in_content(text, model=model, timeout=timeout, temperature=temperature)
         print(f"Y/N Initial Answer: {answer}") if self.verbose == True else 0
-            
+
         loops = 0
         while loops < loop_limit and answer not in ['yes', 'no']:
             answer = self.isolate_yes_no(answer, timeout=timeout)
@@ -153,7 +153,7 @@ User: The following content should be a number - Content: "{content}"
 
     def get_match(self, text: str, list_of_options: list, model=None, loop_limit=4, timeout=None, temperature=None) -> str:
         '''
-        This function can be used in a variety of Natural Language Processing (NLP) tasks, 
+        This function can be used in a variety of Natural Language Processing (NLP) tasks,
         such as text classification or intent recognition.
 
         Classify any text input to a single option contained in a list of options. - Returns exact match to one of items on list.
@@ -163,7 +163,7 @@ User: The following content should be a number - Content: "{content}"
         list_copy = []
         for option in list_of_options:
             list_copy.append(str(option).strip().replace('.', '').lower().strip('"').strip("'"))
-        prompt = f"""Respond with one of the options on this list: {list_copy} 
+        prompt = f"""Respond with one of the options on this list: {list_copy}
 Content to classify: "{text}"  \n\nDo not respond with anything other than the option on the list. Do not add any additional spaces or characters other than the label: {list_copy}"""
 
         answer = self.retry_llm(prompt, model, loop_limit, timeout=timeout, temperature=temperature)
@@ -186,18 +186,18 @@ Content to classify: "{text}"  \n\nDo not respond with anything other than the o
                     pass
 
         return new_answer
-    
+
 
     def get_multi_match(self, text: str, list_of_options: list, model=None, loop_limit=4, timeout=None, temperature=None) -> list:
-        prompt = f"""Respond with a list of items from this list: {list_of_options} 
+        prompt = f"""Respond with a list of items from this list: {list_of_options}
         Content to classify: "{text}"
-        
+
         Respond with a subset of matches from the list above. Each match should be on a new line, without any additional formatting or punctuation."""
 
         answer = self.retry_llm(prompt, model, loop_limit, timeout=timeout, temperature=temperature)
         print(f'Get Answer: {answer}') if self.verbose else None
 
-        # Split the answer into lines and strip whitespace
+        answer = answer.replace(',', '\n').replace(';', '\n')
         candidates = [line.strip() for line in answer.split('\n') if line.strip()]
 
         # Process the candidates, directly adding matches, and using get_match for others
@@ -221,13 +221,13 @@ Content to classify: "{text}"  \n\nDo not respond with anything other than the o
         list_copy = []
         for option in list_of_options:
             list_copy.append(option.strip().replace('.', '').lower().strip('"').strip("'"))
-        prompt_template = """Respond with one of the options on this list: {list_of_options} 
+        prompt_template = """Respond with one of the options on this list: {list_of_options}
 Content to classify: "{content}"  \n\nClassifiy the content above based on which topic it is mostly related to one topic: {list_of_options}"""
-        
+
         prompt = prompt_template.format(content=text, list_of_options=list_copy)
         topic = self.retry_llm(prompt, model, loop_limit, timeout=timeout, temperature=temperature)
         print(f"Topic Answer: {topic}") if self.verbose else 0
-            
+
         if topic is not None:
             topic = list_of_options[list_copy.index(topic)]  # return original topic
 
@@ -235,25 +235,25 @@ Content to classify: "{content}"  \n\nClassifiy the content above based on which
 
 
     def match_or_make(self, text, list_of_options: list = [], model=None, loop_limit=20) -> str:
-        ''' 
-            "M&M" Returns exact match or new option. Input text and list_of_options. 
+        '''
+            "M&M" Returns exact match or new option. Input text and list_of_options.
             If no list is input, then one will be made and be output
 
-            Internally, decides whether or not to make a new option. 
+            Internally, decides whether or not to make a new option.
             Then decides on match is no new option.
             If new option, create it, then finalize it
         '''
         if not list_of_options:
             answer = self.make_option_from_zero(text, list_of_options)
         else:
-            yn_prompt = f'''Is this content able to be categorized in one of the provided categories? Categories: {list_of_options} 
-Respond "No" if a new category should be made. 
+            yn_prompt = f'''Is this content able to be categorized in one of the provided categories? Categories: {list_of_options}
+Respond "No" if a new category should be made.
 Respond "Yes" if the right category already exists in the list'''
             binary = self.get_yes_no(text, yn_prompt, model=model)
-            
+
             if binary == 'yes':
                 answer = self.get_match(text, list_of_options)
-            
+
             if binary == 'no':
                 answer = self.make_option(text, list_of_options)
 
@@ -273,11 +273,11 @@ Respond "Yes" if the right category already exists in the list'''
                 loops += 1
                 if (len(answer.split())) <= 2:
                     break
-        
+
         return answer
-    
 
-    # internal function 
+
+    # internal function
     def make_option(self, text, list_of_options: list, model=None, timeout=None, temperature=None) -> str:
         prompt_template = """Content to classify: "{content}"  \n\n
 Create a new category for the content based on these other categories in this list: {list_of_options}"""
@@ -287,35 +287,35 @@ Create a new category for the content based on these other categories in this li
         return self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
 
 
-    # internal function 
+    # internal function
     def make_option_from_zero(self, text, model=None, timeout=None, temperature=None):
         '''
             Option without anything to go off of but the text input
         '''
-        prompt_template = """Given the following content: {content} 
-Make a simple and general category to classify content like this. Respond only with the category, and 
+        prompt_template = """Given the following content: {content}
+Make a simple and general category to classify content like this. Respond only with the category, and
 no explination before or after, just the name of the catagory. \n\nThe name of the category is: """
-        
+
         prompt = prompt_template.format(content=text)
 
         return self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
-    
 
-    # internal function 
+
+    # internal function
     def finalize_category(self, text, prev_answer, timeout=None, temperature=None):
         prompt_template = """Given the following category suggestion: "{prev_answer}"
 \nand this text: "{text}" The final and simple one or two word category name I created for it is: """
-        
+
         prompt = prompt_template.format(prev_answer=prev_answer, text=text)
 
         return self.retry_llm(custom_prompt=prompt, timeout=timeout)
 
 
-    # internal function 
+    # internal function
     def isolate_yes_no(self, content, question: str, model=None, timeout=None, temperature=None):
         '''Not recommended for external use. Internal function'''
-        prompt_template = """Do not respond with anything before the yes or no. Do not add anything after the "yes" or "no". 
-Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: 'Yes' Example question 2: 'Do you think I am fat?' 
+        prompt_template = """Do not respond with anything before the yes or no. Do not add anything after the "yes" or "no".
+Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: 'Yes' Example question 2: 'Do you think I am fat?'
 Example answer 2: 'No' Example question 3: 'Should I use Matplotlib in Python to draw a graph of csv information I have' Example answer 3: 'Yes'
 Example question 4: 'As an ai language model, I can't tell you yes or no, but I it does have a tendency to work sometimes.' Example answer 4: 'Yes'
 Example question 5: 'Will this happen if it's 19 percent likely to happen?' Example answer 5: 'No'
@@ -324,13 +324,13 @@ Example question 5: 'Will this happen if it's 19 percent likely to happen?' Exam
         prompt = prompt_template.format(content=content, question=question)
 
         return self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
-    
 
-    # internal function 
+
+    # internal function
     def yay_or_nay(self, content, question: str, model=None, timeout=None, temperature=None):
         '''Not recommended for external use. Internal function'''
-        prompt_template = """Do not respond with anything before the yes or no. Do not add anything after the "yes" or "no". 
-Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: 'Yes' Example question 2: 'Do you think I am fat?' 
+        prompt_template = """Do not respond with anything before the yes or no. Do not add anything after the "yes" or "no".
+Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: 'Yes' Example question 2: 'Do you think I am fat?'
 Example answer 2: 'No' Example question 3: 'Should I use Matplotlib in Python to draw a graph of csv information I have' Example answer 3: 'Yes'
 Example question 4: 'As an ai language model, I can't tell you yes or no, but I it does have a tendency to work sometimes.' Example answer 4: 'Yes'
 Example question 5: 'Will this happen if it's 19 percent likely to happen?' Example answer 5: 'No'
@@ -339,31 +339,31 @@ Example question 5: 'Will this happen if it's 19 percent likely to happen?' Exam
         prompt = prompt_template.format(content=content, question=question)
 
         return self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
-    
 
-    # internal function 
+
+    # internal function
     def yay_or_nay_question_in_content(self, content: str, model=None, timeout=None, temperature=None):
         '''Not recommended for external use. Internal function'''
-        prompt_template = """Do not respond with anything before the yes or no. Do not add anything after the "yes" or "no". 
-Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: 'Yes' Example question 2: 'Do you think I am fat?' 
+        prompt_template = """Do not respond with anything before the yes or no. Do not add anything after the "yes" or "no".
+Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: 'Yes' Example question 2: 'Do you think I am fat?'
 Example answer 2: 'No' Example question 3: 'Should I use Matplotlib in Python to draw a graph of csv information I have' Example answer 3: 'Yes'
 Example question 4: 'As an ai language model, I can't tell you yes or no, but I it does have a tendency to work sometimes.' Example answer 4: 'Yes'
 Example question 5: 'Will this happen if it's 19 percent likely to happen?' Example answer 5: 'No'
 User: "{content}"
 
-Respond with a "yes" or "no" 
+Respond with a "yes" or "no"
 Agent:"""
 
         prompt = prompt_template.format(content=content)
 
         return self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
-    
 
-    # internal function 
+
+    # internal function
     def zero_or_one(self, content, zero_if: str, one_if: str, model=None, timeout=None, temperature=None):
         '''Not recommended for external use. Internal function'''
-        prompt_template = """Do not respond with anything before the 1 or 0. Do not add anything after the "1" or "0". 
-Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: '1' Example question 2: 'Do you think I am fat?' 
+        prompt_template = """Do not respond with anything before the 1 or 0. Do not add anything after the "1" or "0".
+Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: '1' Example question 2: 'Do you think I am fat?'
 Example answer 2: '0' Example question 3: 'Should I use Matplotlib in Python to draw a graph of csv information I have' Example answer 3: '1'
 Example question 4: 'As an ai language model, I can't tell you yes or no, but I it does have a tendency to work sometimes.' Example answer 4: '1'
 Example question 5: 'Will this happen if it's 19 percent likely to happen?' Example answer 5: '0'
@@ -372,13 +372,13 @@ Example question 5: 'Will this happen if it's 19 percent likely to happen?' Exam
         prompt = prompt_template.format(content=content, zero_if=zero_if, one_if=one_if)
 
         return self.retry_llm(custom_prompt=prompt, model=model, timeout=timeout, temperature=temperature)
-    
 
-    # internal function 
+
+    # internal function
     def isolate_zero_one(self, content, model=None, timeout=None, temperature=None):
         '''Not recommended for external use. Internal function'''
-        prompt_template = """Do not respond with anything before the 1 or 0. Do not add anything after the "1" or "0". 
-Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: '1' Example question 2: 'Do you think I am fat?' 
+        prompt_template = """Do not respond with anything before the 1 or 0. Do not add anything after the "1" or "0".
+Example question 1: 'Is the Eiffel tower located in Paris?' Example answer 1: '1' Example question 2: 'Do you think I am fat?'
 Example answer 2: '0' Example question 3: 'Should I use Matplotlib in Python to draw a graph of csv information I have' Example answer 3: '1'
 Example question 4: 'As an ai language model, I can't tell you yes or no, but I it does have a tendency to work sometimes.' Example answer 4: '1'
 Example question 5: 'Will this happen if it's 19 percent likely to happen?' Example answer 5: '0'
@@ -427,9 +427,9 @@ Example question 5: 'Will this happen if it's 19 percent likely to happen?' Exam
 
     def perfect(self, content, instructions: str, model=None, timeout=None, temperature=None, yes_no_timeout=None):
         '''Wrapper function ensures that you get exactly what you wanted, and will not return until you get what you wanted'''
-        output = self.get_yes_no(model=model, text=f"These are the instructions: {instructions} and this is the Output: {content}", 
+        output = self.get_yes_no(model=model, text=f"These are the instructions: {instructions} and this is the Output: {content}",
                     question=f"Does the output match the instructions?", timeout=yes_no_timeout) == 'no'
-        
+
         while output:
             new_intstructions = f'''The following content does not match the instructions exactly. Your job is to return the content exactly as the instructions direct:
 Instructions: {instructions}
@@ -437,10 +437,10 @@ Instructions: {instructions}
 Content: {content}
 
 Return the content exactly as the instructions direct'''
-            
+
             content = self.llm(new_intstructions, model=model, timeout=timeout, temperature=temperature)
             if self.verbose == True:
                 print("Content:", content)
-            output = self.get_yes_no(model=model, text=f"These are the instructions: {instructions}", 
+            output = self.get_yes_no(model=model, text=f"These are the instructions: {instructions}",
                         question=f"This is the Output: {content} \n\nDoes this output match the instructions?", timeout=yes_no_timeout) == 'no'
         return content