Spaces:
Paused
Paused
Added logging with the logging package
Browse files
app.py
CHANGED
|
@@ -15,6 +15,9 @@ from sentence_transformers import SentenceTransformer
|
|
| 15 |
from peft import PeftModel
|
| 16 |
from bs4 import BeautifulSoup
|
| 17 |
import requests
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
headers = {
|
| 20 |
"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_5) AppleWebKit 537.36 (KHTML, like Gecko) Chrome",
|
|
@@ -24,7 +27,7 @@ headers = {
|
|
| 24 |
|
| 25 |
|
| 26 |
def google_search(text):
|
| 27 |
-
|
| 28 |
try:
|
| 29 |
site = requests.get(f"https://www.google.com/search?hl=en&q={text}", headers=headers)
|
| 30 |
main = (
|
|
@@ -40,10 +43,10 @@ def google_search(text):
|
|
| 40 |
|
| 41 |
ans = " \n".join(res)
|
| 42 |
except Exception as ex:
|
| 43 |
-
|
| 44 |
ans = ""
|
| 45 |
|
| 46 |
-
|
| 47 |
|
| 48 |
return ans
|
| 49 |
|
|
@@ -112,13 +115,13 @@ def get_results_from_pinecone(query, top_k=3, re_rank=True, verbose=True):
|
|
| 112 |
return []
|
| 113 |
|
| 114 |
if verbose:
|
| 115 |
-
|
| 116 |
|
| 117 |
final_results = []
|
| 118 |
|
| 119 |
if re_rank:
|
| 120 |
if verbose:
|
| 121 |
-
|
| 122 |
|
| 123 |
sentence_combinations = [
|
| 124 |
[query, result_from_pinecone["metadata"]["text"]] for result_from_pinecone in results_from_pinecone
|
|
@@ -135,17 +138,17 @@ def get_results_from_pinecone(query, top_k=3, re_rank=True, verbose=True):
|
|
| 135 |
result_from_pinecone = results_from_pinecone[idx]
|
| 136 |
final_results.append(result_from_pinecone)
|
| 137 |
if verbose:
|
| 138 |
-
|
| 139 |
f"{result_from_pinecone['id']}\t{result_from_pinecone['score']:.2f}\t{similarity_scores[idx]:.2f}\t{result_from_pinecone['metadata']['text'][:50]}"
|
| 140 |
)
|
| 141 |
return final_results
|
| 142 |
|
| 143 |
if verbose:
|
| 144 |
-
|
| 145 |
for result_from_pinecone in results_from_pinecone:
|
| 146 |
final_results.append(result_from_pinecone)
|
| 147 |
if verbose:
|
| 148 |
-
|
| 149 |
f"{result_from_pinecone['id']}\t{result_from_pinecone['score']:.2f}\t{result_from_pinecone['metadata']['text'][:50]}"
|
| 150 |
)
|
| 151 |
|
|
@@ -268,13 +271,13 @@ def text_to_text_generation(verbose, prompt):
|
|
| 268 |
match response_num:
|
| 269 |
case 0:
|
| 270 |
prompt = f"[INST] {prompt}\n Lets think step by step. [/INST] {start_template}"
|
| 271 |
-
|
| 272 |
-
|
| 273 |
case 1:
|
| 274 |
if retriever == "semantic_search":
|
| 275 |
question = prompt
|
| 276 |
-
|
| 277 |
-
|
| 278 |
(
|
| 279 |
f"You are a helpful kubernetes professional. [INST] Use the following documentation, if it is relevant to answer the question below. [/INST]\nDocumentation: [RETRIEVED_RESULTS_FROM_BOOK] [INST] Answer the following question: {question} [/INST]\nAnswer: \n")
|
| 280 |
|
|
@@ -295,8 +298,8 @@ def text_to_text_generation(verbose, prompt):
|
|
| 295 |
question = prompt
|
| 296 |
prompt = f"You are a helpful kubernetes professional. [INST] Use the following documentation, if it is relevant to answer the question below. [/INST]\nDocumentation: {retrieved_results} </s>\n<s> [INST] Answer the following question: {prompt} [/INST]\nAnswer: "
|
| 297 |
|
| 298 |
-
|
| 299 |
-
|
| 300 |
(
|
| 301 |
f"You are a helpful kubernetes professional. [INST] Use the following documentation, if it is relevant to answer the question below. [/INST]\nDocumentation: [RETRIEVED_RESULTS_FROM_GOOGLE] [INST] Answer the following question: {question} [/INST]\nAnswer:\n\n"
|
| 302 |
)
|
|
@@ -312,13 +315,13 @@ def text_to_text_generation(verbose, prompt):
|
|
| 312 |
)
|
| 313 |
else:
|
| 314 |
prompt = f"[INST] Answer the following question: {prompt} [/INST]\nAnswer: "
|
| 315 |
-
|
| 316 |
-
|
| 317 |
|
| 318 |
case _:
|
| 319 |
prompt = f"[INST] {prompt} [/INST]"
|
| 320 |
-
|
| 321 |
-
|
| 322 |
|
| 323 |
return prompt, md
|
| 324 |
|
|
@@ -350,8 +353,8 @@ def text_to_text_generation(verbose, prompt):
|
|
| 350 |
|
| 351 |
modes = ["Kubectl command", "Kubernetes related", "Other"]
|
| 352 |
|
| 353 |
-
|
| 354 |
-
|
| 355 |
|
| 356 |
modes[response_num] = f"**{modes[response_num]}**"
|
| 357 |
modes = " / ".join(modes)
|
|
@@ -419,16 +422,16 @@ def text_to_text_generation(verbose, prompt):
|
|
| 419 |
res_prompt, res_semantic_search_prompt, res_google_search_prompt
|
| 420 |
)
|
| 421 |
|
| 422 |
-
|
| 423 |
-
|
| 424 |
|
| 425 |
|
| 426 |
res_prompt, res_normal = cleanup(*gen_normal)
|
| 427 |
res_semantic_search_prompt, res_semantic_search = cleanup(*gen_semantic_search)
|
| 428 |
res_google_search_prompt, res_google_search = cleanup(*gen_google_search)
|
| 429 |
|
| 430 |
-
|
| 431 |
-
|
| 432 |
|
| 433 |
if verbose:
|
| 434 |
return (
|
|
|
|
| 15 |
from peft import PeftModel
|
| 16 |
from bs4 import BeautifulSoup
|
| 17 |
import requests
|
| 18 |
+
import logging
|
| 19 |
+
|
| 20 |
+
logging.basicConfig(format='[%(asctime)s] %(message)s', datefmt='%d-%b-%y %H:%M:%S', level=logging.INFO)
|
| 21 |
|
| 22 |
headers = {
|
| 23 |
"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_5) AppleWebKit 537.36 (KHTML, like Gecko) Chrome",
|
|
|
|
| 27 |
|
| 28 |
|
| 29 |
def google_search(text):
|
| 30 |
+
logging.info(f"Google search on: {text}")
|
| 31 |
try:
|
| 32 |
site = requests.get(f"https://www.google.com/search?hl=en&q={text}", headers=headers)
|
| 33 |
main = (
|
|
|
|
| 43 |
|
| 44 |
ans = " \n".join(res)
|
| 45 |
except Exception as ex:
|
| 46 |
+
logging.error(f"Error: {ex}")
|
| 47 |
ans = ""
|
| 48 |
|
| 49 |
+
logging.info(f"The result of the google search is: {ans}")
|
| 50 |
|
| 51 |
return ans
|
| 52 |
|
|
|
|
| 115 |
return []
|
| 116 |
|
| 117 |
if verbose:
|
| 118 |
+
logging.info("Query:", query)
|
| 119 |
|
| 120 |
final_results = []
|
| 121 |
|
| 122 |
if re_rank:
|
| 123 |
if verbose:
|
| 124 |
+
logging.info("Document ID (Hash)\t\tRetrieval Score\tCE Score\tText")
|
| 125 |
|
| 126 |
sentence_combinations = [
|
| 127 |
[query, result_from_pinecone["metadata"]["text"]] for result_from_pinecone in results_from_pinecone
|
|
|
|
| 138 |
result_from_pinecone = results_from_pinecone[idx]
|
| 139 |
final_results.append(result_from_pinecone)
|
| 140 |
if verbose:
|
| 141 |
+
logging.info(
|
| 142 |
f"{result_from_pinecone['id']}\t{result_from_pinecone['score']:.2f}\t{similarity_scores[idx]:.2f}\t{result_from_pinecone['metadata']['text'][:50]}"
|
| 143 |
)
|
| 144 |
return final_results
|
| 145 |
|
| 146 |
if verbose:
|
| 147 |
+
logging.info("Document ID (Hash)\t\tRetrieval Score\tText")
|
| 148 |
for result_from_pinecone in results_from_pinecone:
|
| 149 |
final_results.append(result_from_pinecone)
|
| 150 |
if verbose:
|
| 151 |
+
logging.info(
|
| 152 |
f"{result_from_pinecone['id']}\t{result_from_pinecone['score']:.2f}\t{result_from_pinecone['metadata']['text'][:50]}"
|
| 153 |
)
|
| 154 |
|
|
|
|
| 271 |
match response_num:
|
| 272 |
case 0:
|
| 273 |
prompt = f"[INST] {prompt}\n Lets think step by step. [/INST] {start_template}"
|
| 274 |
+
logging.info('Kubectl command prompt:')
|
| 275 |
+
logging.info(prompt)
|
| 276 |
case 1:
|
| 277 |
if retriever == "semantic_search":
|
| 278 |
question = prompt
|
| 279 |
+
logging.info('Semantic search prompt:')
|
| 280 |
+
logging.info(
|
| 281 |
(
|
| 282 |
f"You are a helpful kubernetes professional. [INST] Use the following documentation, if it is relevant to answer the question below. [/INST]\nDocumentation: [RETRIEVED_RESULTS_FROM_BOOK] [INST] Answer the following question: {question} [/INST]\nAnswer: \n")
|
| 283 |
|
|
|
|
| 298 |
question = prompt
|
| 299 |
prompt = f"You are a helpful kubernetes professional. [INST] Use the following documentation, if it is relevant to answer the question below. [/INST]\nDocumentation: {retrieved_results} </s>\n<s> [INST] Answer the following question: {prompt} [/INST]\nAnswer: "
|
| 300 |
|
| 301 |
+
logging.info('Google search prompt:')
|
| 302 |
+
logging.info(
|
| 303 |
(
|
| 304 |
f"You are a helpful kubernetes professional. [INST] Use the following documentation, if it is relevant to answer the question below. [/INST]\nDocumentation: [RETRIEVED_RESULTS_FROM_GOOGLE] [INST] Answer the following question: {question} [/INST]\nAnswer:\n\n"
|
| 305 |
)
|
|
|
|
| 315 |
)
|
| 316 |
else:
|
| 317 |
prompt = f"[INST] Answer the following question: {prompt} [/INST]\nAnswer: "
|
| 318 |
+
logging.info('No retriever question prompt:')
|
| 319 |
+
logging.info(prompt)
|
| 320 |
|
| 321 |
case _:
|
| 322 |
prompt = f"[INST] {prompt} [/INST]"
|
| 323 |
+
logging.info('Other question prompt:')
|
| 324 |
+
logging.info(prompt)
|
| 325 |
|
| 326 |
return prompt, md
|
| 327 |
|
|
|
|
| 353 |
|
| 354 |
modes = ["Kubectl command", "Kubernetes related", "Other"]
|
| 355 |
|
| 356 |
+
logging.info(f'{" Query Start ":-^40}')
|
| 357 |
+
logging.info("Classified as: " + modes[response_num])
|
| 358 |
|
| 359 |
modes[response_num] = f"**{modes[response_num]}**"
|
| 360 |
modes = " / ".join(modes)
|
|
|
|
| 422 |
res_prompt, res_semantic_search_prompt, res_google_search_prompt
|
| 423 |
)
|
| 424 |
|
| 425 |
+
logging.info("SEMANTIC BEFORE CLEANUP: ", gen_semantic_search)
|
| 426 |
+
logging.info("GOOGLE BEFORE CLEANUP: ", gen_google_search)
|
| 427 |
|
| 428 |
|
| 429 |
res_prompt, res_normal = cleanup(*gen_normal)
|
| 430 |
res_semantic_search_prompt, res_semantic_search = cleanup(*gen_semantic_search)
|
| 431 |
res_google_search_prompt, res_google_search = cleanup(*gen_google_search)
|
| 432 |
|
| 433 |
+
logging.info("SEMANTIC AFTER CLEANUP: ", res_semantic_search)
|
| 434 |
+
logging.info("GOOGLE AFTER CLEANUP: ", res_google_search)
|
| 435 |
|
| 436 |
if verbose:
|
| 437 |
return (
|