Update app.py

app.py

@@ -48,7 +48,7 @@ def parse_action(string: str):
 
 VERBOSE = True
 MAX_HISTORY = 100
-MAX_DATA = 1000
+MAX_DATA = 20000
 
 def format_prompt(message, history):
   prompt = "<s>"
@@ -58,61 +58,6 @@ def format_prompt(message, history):
   prompt += f"[INST] {message} [/INST]"
   return prompt
 
-def call_search(purpose, task, history, action_input):
-    return_list=[]
-    print (action_input)
-    #if action_input in query.tasks:
-    print ("trying")        
-    try:
-        if action_input != "" and action_input != None:
-            action_input.strip('""')
-            #model_list = api.list_models(filter=f"{action_input}",sort="last_modified",limit=1000,direction=-1)
-            #model_list = api.list_models(filter=f"{action_input}",limit=1000)
-            model_list = api.list_models(filter=f"{action_input}")
-            this_obj = list(model_list)
-            print(f'THIS_OBJ :: {this_obj[0]}')
-            for i,eb in enumerate(this_obj):
-                #return_list.append(this_obj[i].id)
-                return_list.append({"id":this_obj[i].id,
-                                    "author":this_obj[i].author,
-                                    "created_at":this_obj[i].created_at,
-                                    "last_modified":this_obj[i].last_modified,
-                                    "private":this_obj[i].private,
-                                    "gated":this_obj[i].gated,
-                                    "disabled":this_obj[i].disabled,
-                                    "downloads":this_obj[i].downloads,
-                                    "likes":this_obj[i].likes,
-                                    "library_name":this_obj[i].library_name,
-                                    "tags":this_obj[i].tags,
-                                    "pipeline_tag":this_obj[i].pipeline_tag,
-                                   })
-            #print (return_list)
-            c=0
-            rl = len(return_list)
-            print(rl)
-            for i in str(return_list):
-                if i == " " or i==",":
-                    c +=1
-            
-            print (c)
-            if rl > MAX_DATA:
-                print("compressing...")
-                return_list = compress_data(rl,purpose,task,return_list)
-            history = "observation: the search results are:\n {}\n".format(return_list)
-            return "MAIN", None, history, task
-        else: 
-            history = "observation: I need to trigger a search using the following syntax:\naction: SEARCH action_input=URL\n"
-            return "UPDATE-TASK", None, history, task
-    except Exception as e:
-        print (e)
-        history = "observation: I need to trigger a search using the following syntax:\naction: SEARCH action_input=URL\n"
-        return "UPDATE-TASK", None, history, task
-
-        #else:
-    #    history = "observation: The search query I used did not return a valid response"
-        
-    return "MAIN", None, history, task
-
 
 def run_gpt(
     prompt_template,
@@ -143,7 +88,7 @@ def run_gpt(
     
     
     #formatted_prompt = format_prompt(f"{system_prompt}, {prompt}", history)
-    #formatted_prompt = format_prompt(f'{content}', history)
+    #formatted_prompt = format_prompt(f'{content}', **prompt_kwargs['history'])
 
     stream = client.text_generation(content, **generate_kwargs, stream=True, details=True, return_full_text=False)
     resp = ""
@@ -155,7 +100,7 @@ def run_gpt(
         print(LOG_RESPONSE.format(resp))
     return resp
 
-def compress_data(c,purpose, task, history):
+def compress_data(c,purpose, task, history, result):
     seed=random.randint(1,1000000000)
     
     print (c)
@@ -180,15 +125,15 @@ def compress_data(c,purpose, task, history):
         hist = history[s:e]
         
         resp = run_gpt(
-            COMPRESS_DATA_PROMPT_SMALL,
+            COMPRESS_DATA_PROMPT,
             stop_tokens=["observation:", "task:", "action:", "thought:"],
-            max_tokens=2048,
+            max_tokens=1024,
             seed=seed,
             purpose=purpose,
             task=task,
             knowledge=new_history,
             history=hist,
-        )
+        ).strip('\n')
         new_history = resp
         print (resp)
         out+=resp
@@ -198,15 +143,16 @@ def compress_data(c,purpose, task, history):
     resp = run_gpt(
         COMPRESS_DATA_PROMPT,
         stop_tokens=["observation:", "task:", "action:", "thought:"],
-        max_tokens=1024,
+        max_tokens=2048,
         seed=seed,
         purpose=purpose,
         task=task,
         knowledge=new_history,
-        history="All data has been recieved.",
-    )'''
+        history=result,
+    )
+    '''
     print ("final" + resp)
-    history = "observation: {}\n".format(resp)
+    history = "result: {}\n".format(resp)
     return history
 
 
@@ -226,41 +172,40 @@ def compress_history(purpose, task, history):
     return history
 
 
-def call_main(purpose, task, history, action_input):
+def call_main(purpose, task, history, action_input, result):
     resp = run_gpt(
         FINDER,
-        stop_tokens=["observation:", "task:", "action:"],
-        max_tokens=512,
+        stop_tokens=["observation:", "task:"],
+        max_tokens=2096,
         seed=random.randint(1,1000000000),
         purpose=purpose,
         task=task,
         history=history,
     )
     lines = resp.strip().strip("\n").split("\n")
+    #history=""
     for line in lines:
         if line == "":
             continue
         if line.startswith("thought: "):
             history += "{}\n".format(line)
-        if line.startswith("action: COMPLETE"):
-            print("COMPLETE called")
-            return "COMPLETE", None, history, task
-        if line.startswith("action:"):
+            
+        if line.startswith("action: "):
             action_name, action_input = parse_action(line)
             print(f'ACTION::{action_name} -- INPUT :: {action_input}')
-            history += "{}\n".format(line)
-            return action_name, action_input, history, task
+            #history += "{}\n".format(line)
+            return action_name, action_input, history, task, result
         else:
-            
-            history += "{}\n".format(line)
+            pass
+            #history += "{}\n".format(line)
             #assert False, "unknown action: {}".format(line)
             #return "UPDATE-TASK", None, history, task
     if "VERBOSE":
         print(history)
-    return "MAIN", None, history, task
+    return "MAIN", None, history, task, result
 
 
-def call_set_task(purpose, task, history, action_input):
+def call_set_task(purpose, task, history, action_input, result):
     task = run_gpt(
         TASK_PROMPT,
         stop_tokens=[],
@@ -271,7 +216,7 @@ def call_set_task(purpose, task, history, action_input):
         history=history,
     ).strip("\n")
     history += "observation: task has been updated to: {}\n".format(task)
-    return "MAIN", None, history, task
+    return "MAIN", None, history, task, result
 
 
 
@@ -282,7 +227,7 @@ def search_all(url):
 
 
 
-def find_all(purpose,task,history, url):
+def find_all(purpose,task,history, url, result):
     return_list=[]
     print (url)
     #if action_input in query.tasks:
@@ -299,89 +244,59 @@ def find_all(purpose,task,history, url):
             # get attributes:
             print(soup.title.name)
             # get values:
-            print(soup.title.string)
+            #print(soup.title.string)
             # beginning navigation:
-            print(soup.title.parent.name)
+            #print(soup.title.parent.name)
             #rawp.append([tag.name for tag in soup.find_all()] )
-            print([tag.name for tag in soup.find_all()])
+            #print([tag.name for tag in soup.find_all()])
             rawp=(f'RAW TEXT RETURNED: {soup.text}')
+            cnt=0
+            cnt+=len(rawp)
             out.append(rawp)
+            out.append("HTML fragments: ")
             q=("a","p","span","content","article")
-            for p in soup.find_all(q):
-                out.append([{q:p.string,"parent":p.parent.name,"previous":[b for b in p.previous],"first-child":[b.name for b in p.children],"content":p}])
+            for p in soup.find_all("a"):
+                #cnt+=len(p.string)
+                #cnt+=len(p)                
+                
+                out.append([{"LINK TITLE":p.get('title'),"URL":p.get('href'),"STRING":p.string}])
             c=0
             out = str(out)
             rl = len(out)
+            #print (cnt)
             print(f'rl:: {rl}')
             #for ea in out:
             for i in str(out):
-                if i == " " or i=="," or i=="\n":
+                if i == " " or i=="," or i=="\n" or i=="/" or i=="." or i=="<":
                     c +=1
             print (f'c:: {c}')
-            if rl > MAX_DATA:
+            if c > MAX_HISTORY:
                 print("compressing...")
-                rawp = compress_data(c,purpose,task,out)    
+                rawp = compress_data(c,purpose,task,out,result)  
+            else:
+                rawp = out 
+            result += rawp
+                
             print (rawp)
             print (f'out:: {out}')
-            history += "observation: the search results are:\n {}\n".format(out)
-            task = "complete?"
-            return "MAIN", None, history, task
+            history += "observation: the search results are:\n {}\n".format(rawp)
+            task = "compile report or complete?"
+            return "MAIN", None, history, task, result
         else: 
-            history += "observation: I need to trigger a search using the following syntax:\naction: SCRAPE_WEBSITE action_input=URL\n"
-            return "MAIN", None, history, task
+            history += "observation: An Error occured\nI need to trigger a search using the following syntax:\naction: SCRAPE_WEBSITE action_input=URL\n"
+            return "MAIN", None, history, task, result
     except Exception as e:
         print (e)
         history += "observation: I need to trigger a search using the following syntax:\naction: SCRAPE_WEBSITE action_input=URL\n"
-        return "MAIN", None, history, task
+        return "MAIN", None, history, task, result
 
         #else:
     #    history = "observation: The search query I used did not return a valid response"
         
-    return "MAIN", None, history, task
+    return "MAIN", None, history, task, result
+
 
 
-def find_it(url,q=None,num=None):
-    out = []
-    out_l = []
-    z=""
-    source = urllib.request.urlopen(url).read()
-    soup = bs4.BeautifulSoup(source,'lxml')
-    
-    for p in soup.find_all(f'{q}'):
-        if num != "":
-            z=p.get(f'{num}')
-        
-        try:
-            test = soup.select(f'{p.name}:first-child')
-            
-            #print(p.findChildren())
-        except Exception as e:
-            print (e)
-        #out.append(p)
-        out.append([{q:p.string,"additional":z,"parent":p.parent.name,"previous":[b for b in p.previous],"first-child":[b.name for b in p.children],"content":p}])
-        if p.string !=None:
-            out_l.append(p.string)
-        else:
-            out_l.append(z)
-        #out.append(p.parent.name)
-        print(dir(p))
-        print(p.parent.name)
-    for url in soup.find_all('a'):
-        print(url.get('href'))
-        
-    #print(soup.get_text())
-    return out,out_l
-    
-def find_it2(url):
-    response = requests.get(url,a1=None,q2=None,q3=None)
-    try:
-        response.raise_for_status()
-        soup = BeautifulSoup(response.content, 'lxml')
-        out = 'URL Links:\n'.join([p.text for p in soup.find_all('a')])
-        return out
-    except Exception as e:
-        print (e)
-        return e
 #################################
 
 NAME_TO_FUNC = {
@@ -392,11 +307,11 @@ NAME_TO_FUNC = {
 }
 
 
-def run_action(purpose, task, history, action_name, action_input):
-    if action_name == "COMPLETE":
+def run_action(purpose, task, history, action_name, action_input,result):
+    if "COMPLETE" in action_name:
         print("Complete - Exiting")
         #exit(0) 
-        return "COMPLETE", None, history, task
+        return "COMPLETE", None, history, task, result
 
     # compress the history when it is long
     if len(history.split("\n")) > MAX_HISTORY:
@@ -408,18 +323,21 @@ def run_action(purpose, task, history, action_name, action_input):
         assert action_name in NAME_TO_FUNC
 
         print(f"RUN: {action_name}  ACTION_INPUT: {action_input}")
-        return NAME_TO_FUNC[action_name](purpose, task, history, action_input)
+        return NAME_TO_FUNC[action_name](purpose, task, history, action_input, result)
     else:
-        history += "observation: The TOOL I tried to use returned an error, I need to select a tool from: (UPDATE-TASK, SEARCH_ENGINE, WEBSITE_SCRAPE, COMPLETE)\n"
+        history += "observation: The TOOL I tried to use returned an error, I need to select a tool from: (UPDATE-TASK, SEARCH_ENGINE, SCRAPE_WEBSITE, COMPLETE)\n"
 
-        return "MAIN", None, history, task
+        return "MAIN", None, history, task, result
 
 def run(purpose,history):
     task=None
-    history = ""
-    #if not history:
-    #    history = []
-    action_name = "SEARCH_ENGINE" if task is None else "MAIN"
+    result=""
+    #history = ""
+    if not history:
+        history = ""
+    else:
+        history=str(history)
+    action_name = "MAIN"
     action_input = None
     while True:
         print("")
@@ -431,36 +349,38 @@ def run(purpose,history):
         #print(history)
         print("---")
 
-        action_name, action_input, history, task = run_action(
+        action_name, action_input, history, task, result = run_action(
             purpose,
             task,
             history,
             action_name,
             action_input,
+            result
         )
-        yield history
+        yield result
         if action_name == "COMPLETE":
-            return history
+            break
+    return result
 
 
 
 
 examples =[
+    "What is the current weather in Florida?",
     "what are todays breaking news stories?",
-    "find the most popular model that I can use to generate an image by providing a text prompt",
-    "return the top 10 models that I can use to identify objects in images",
-    "which models have the most likes from each category?"
+    "Find the best deals on flippers for scuba diving",
+    "Teach me to fly a helicopter"
 ]
 
 
-app = gr.ChatInterface(
+gr.ChatInterface(
     fn=run,
     chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, likeable=True, layout="panel"),
     title="Mixtral 46.7B Powered <br> Search",
     examples=examples,
     concurrency_limit=20,
-)
-app.launch()
+    css="height:600px;"
+).launch()
 '''
 with gr.Blocks() as app:
     with gr.Row():
@@ -483,4 +403,4 @@ with gr.Blocks() as app:
     
 app.launch()
 
-'''
+'''