Spaces:
Sleeping
Sleeping
Update prompts.py
Browse files- prompts.py +5 -8
prompts.py
CHANGED
@@ -15,8 +15,8 @@ Search Purpose:
|
|
15 |
FINDER = """
|
16 |
|
17 |
Instructions
|
18 |
-
- Use
|
19 |
-
- Use
|
20 |
- Find the pertinent information in the text that you scrape
|
21 |
- When you are finished, return with action: COMPLETE
|
22 |
|
@@ -24,23 +24,20 @@ Use the following format:
|
|
24 |
task: choose the next action from your available tools
|
25 |
action: the action to take (should be one of [UPDATE-TASK, SEARCH_ENGINE, WEBSITE_SCRAPE, COMPLETE]) action_input=XXX
|
26 |
observation: the result of the action
|
27 |
-
action: SEARCH action_input=
|
28 |
action: COMPLETE
|
29 |
|
30 |
Example:
|
31 |
***************************
|
32 |
User command: Find me the breaking news from today
|
33 |
action: SEARCH_ENGINE action_input=https://www.google.com/search?q=todays+breaking+news
|
34 |
-
observation: the data that is returned including urls to websites
|
35 |
-
action: SCRAPE_WEBSITE action_input=https://www.news-website.com
|
36 |
-
observation: the data that is returned
|
37 |
|
38 |
Response:
|
39 |
Assistant: I found the the following news articles.....
|
40 |
action: COMPLETE
|
41 |
***************************
|
42 |
-
|
43 |
-
|
44 |
{history}"""
|
45 |
|
46 |
MODEL_FINDER_PRE = """
|
|
|
15 |
FINDER = """
|
16 |
|
17 |
Instructions
|
18 |
+
- Use the provided tool to find a website to scrape
|
19 |
+
- Use the tool provided tool to scrape the text from the website url
|
20 |
- Find the pertinent information in the text that you scrape
|
21 |
- When you are finished, return with action: COMPLETE
|
22 |
|
|
|
24 |
task: choose the next action from your available tools
|
25 |
action: the action to take (should be one of [UPDATE-TASK, SEARCH_ENGINE, WEBSITE_SCRAPE, COMPLETE]) action_input=XXX
|
26 |
observation: the result of the action
|
27 |
+
action: SEARCH action_input=URL
|
28 |
action: COMPLETE
|
29 |
|
30 |
Example:
|
31 |
***************************
|
32 |
User command: Find me the breaking news from today
|
33 |
action: SEARCH_ENGINE action_input=https://www.google.com/search?q=todays+breaking+news
|
|
|
|
|
|
|
34 |
|
35 |
Response:
|
36 |
Assistant: I found the the following news articles.....
|
37 |
action: COMPLETE
|
38 |
***************************
|
39 |
+
|
40 |
+
Progress:
|
41 |
{history}"""
|
42 |
|
43 |
MODEL_FINDER_PRE = """
|