Experimental async requests

main.py

@@ -26,6 +26,8 @@ import os
 import pytesseract
 import lang
 
+import httpx
+
 from secrets import SystemRandom
 
 from random import randint, sample
@@ -184,19 +186,20 @@ async def __query_ml_predict(qtype: QType, content: str, header: str, token_limi
         case QType.WH:
             
             # Make request to Awan LLM endpoint
-            _r = requests.post(
-                url="https://api.awanllm.com/v1/chat/completions",
-                headers={'Content-Type': 'application/json', 'Authorization': f'Bearer {LLM_API_KEY}'},
-                data=json.dumps({
-                    "model": "Meta-Llama-3-8B-Instruct",
-                    "messages": [
-                        {"role": "user", "content": prompt.gen_prompt_wh(content=content, header=header, num_qs=num_qs, lang=l)}
-                    ],
-                    "max_tokens": max(token_limit, 4096),
-                    "presence_penalty":0.3,
-                    "temperature":0.55
-                })
-            )
+            async with httpx.AsyncClient() as client:
+                _r = await client.post(
+                    url="https://api.awanllm.com/v1/chat/completions",
+                    headers={'Content-Type': 'application/json', 'Authorization': f'Bearer {LLM_API_KEY}'},
+                    data=json.dumps({
+                        "model": "Meta-Llama-3-8B-Instruct",
+                        "messages": [
+                            {"role": "user", "content": prompt.gen_prompt_wh(content=content, header=header, num_qs=num_qs, lang=l)}
+                        ],
+                        "max_tokens": max(token_limit, 4096),
+                        "presence_penalty":0.3,
+                        "temperature":0.55
+                    })
+                )
 
             print(time() - stopwatch)
             return {"content": _r.json()['choices'][0]['message']['content'], "style": QType.WH}
@@ -204,32 +207,33 @@ async def __query_ml_predict(qtype: QType, content: str, header: str, token_limi
         case QType.STMT:
 
             # Make request to Awan LLM endpoint
-            _r = requests.post(
-                url="https://api.awanllm.com/v1/chat/completions",
-                headers={'Content-Type': 'application/json', 'Authorization': f'Bearer {LLM_API_KEY}'},
-                data=json.dumps({
-                    "model": "Meta-Llama-3-8B-Instruct",
-                    "messages": [
-                        {"role": "user", "content": prompt.gen_prompt_statements(content=content, header=header, num_qs=num_qs, lang=l)}
-                    ],
-                    "max_tokens": max(token_limit, 4096),
-                    
-                })
-            )
+            async with httpx.AsyncClient() as client:
+                _r = await client.post(
+                    url="https://api.awanllm.com/v1/chat/completions",
+                    headers={'Content-Type': 'application/json', 'Authorization': f'Bearer {LLM_API_KEY}'},
+                    data=json.dumps({
+                        "model": "Meta-Llama-3-8B-Instruct",
+                        "messages": [
+                            {"role": "user", "content": prompt.gen_prompt_statements(content=content, header=header, num_qs=num_qs, lang=l)}
+                        ],
+                        "max_tokens": max(token_limit, 4096),
+                        
+                    })
+                )
 
             _r_content = _r.json()['choices'][0]['message']['content'].split('\n\n',1)[1]
-
-            _w = requests.post(
-                url="https://api.awanllm.com/v1/chat/completions",
-                headers={'Content-Type': 'application/json', 'Authorization': f'Bearer {LLM_API_KEY}'},
-                data=json.dumps({
-                    "model": "Meta-Llama-3-8B-Instruct",
-                    "messages": [
-                        {"role": "user", "content": prompt.gen_prompt_statements_false(content=_r_content, lang=l)}
-                    ],
-                    "max_tokens": max(token_limit, 4096),
-                    
-                })
+            async with httpx.AsyncClient() as client:
+                _w = await client.post(
+                    url="https://api.awanllm.com/v1/chat/completions",
+                    headers={'Content-Type': 'application/json', 'Authorization': f'Bearer {LLM_API_KEY}'},
+                    data=json.dumps({
+                        "model": "Meta-Llama-3-8B-Instruct",
+                        "messages": [
+                            {"role": "user", "content": prompt.gen_prompt_statements_false(content=_r_content, lang=l)}
+                        ],
+                        "max_tokens": max(token_limit, 4096),
+                        
+                    })
             )
 
             _w_content = _w.json()['choices'][0]['message']['content'].split('\n\n',1)[1]