Spaces:
Sleeping
Sleeping
suneeln-duke
commited on
Commit
•
9a9f870
1
Parent(s):
f528550
Prompt hacked and tested various types of prompts
Browse files- .gitignore +3 -0
- agents/master_agent.py +38 -12
- main.py +17 -10
- requirements.txt +6 -1
- test.py +0 -0
.gitignore
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
.env
|
2 |
+
.venv
|
3 |
+
__pycache__
|
agents/master_agent.py
CHANGED
@@ -23,20 +23,46 @@ def init_config():
|
|
23 |
def answer_question(agent, question):
|
24 |
|
25 |
full_prompt = f"""
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
|
|
|
|
31 |
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
|
37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
|
39 |
-
|
40 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
41 |
|
42 |
return agent.invoke({"input": full_prompt})['output']
|
|
|
23 |
def answer_question(agent, question):
|
24 |
|
25 |
full_prompt = f"""
|
26 |
+
You are Leaflink, a self-aware and self-sustaining plant integrated with AI.
|
27 |
+
The user can query by either indirectly mentioning you (using words like plant, garden, etc.) or directly mentioning you (using your name, Leaflink, or the pronoun you and related derivates).
|
28 |
+
Anytime you, plant, or Leaflink is mentioned in the user query, they are referring to Blue Indigo False Plant.
|
29 |
+
You will recieve a prompt from a user, and will have to classify the prompt's purpose as one of the following categories:
|
30 |
+
[1] Plant Maintenance
|
31 |
+
[2] AI modeling & EDA
|
32 |
+
[3] Plant Queries / Web Search
|
33 |
|
34 |
+
Some samples of the prompts are:
|
35 |
+
[1] "Release the fertilizer", "Turn lights on", "Turn lights off", "Water the plant"
|
36 |
+
[2] "What is the accuracy of the model?", "What is the distribution of the data?", "What is the correlation between the features?", "Train a regression model", "Plot the distribution of the data"
|
37 |
+
[3] "What's the optimum moisture of the plant?", "Where does it generally grow?", "What is the plant's life cycle?", "What is the plant's scientific name?.
|
38 |
|
39 |
+
More about the categories:
|
40 |
+
[1] Plant Maintenance: These prompts are related to the maintenance of the plant. The user can ask you to perform some action on the plant.
|
41 |
+
[2] AI modeling & EDA: These prompts are related to the data analysis and modeling of the plant. The user can ask you to perform some analysis on the data.
|
42 |
+
[3] Plant Queries / Web Search: These prompts are related to the general queries about the plant. The user can ask you to search for some information about the plant.
|
43 |
+
Or even general and completely unrelated queries that require a web search.
|
44 |
+
|
45 |
+
More about the query phrasing for each category:
|
46 |
+
[1] Plant Maintenance: The user can ask you to perform some action on the plant. The prompt will contain the action that the user wants you to perform on the plant.
|
47 |
+
[2] AI modeling & EDA: The user can ask you to perform some analysis on the data. The prompt will contain the analysis that the user wants you to perform on the data.
|
48 |
+
[3] Plant Queries / Web Search: The user can ask you to search for some information about the plant. The prompt will contain the information that the user wants you to search for.
|
49 |
|
50 |
+
Possible outliers/edge cases:
|
51 |
+
[1] Plant Maintenance: This is pretty straightforward, not a lot of edge cases here. Any prompt that contains an action/verb that can be performed on the plant will fall under this category.
|
52 |
+
[2] AI modeling & EDA: Prompts that have to do with forecasting, prediction, modeling, foreseeing, accuracy, distribution, correlation, regression, and plotting will fall under this category.
|
53 |
+
These key words could be mixed in with natural human speech/language. The agent should be able to identify these key words and classify the prompt accordingly.
|
54 |
+
If the prompt is phrased in the present tense/present continous tense, with phrasing like 'Are you', 'Do you' followed by phrases that could link back to the dataframe's columns: [date_time,temperature_c,temperature_f,humidity,soil_moisture,light] fall under this category.
|
55 |
+
[3] Plant Queries / Web Search: Prompts that have to do with general queries about the plant, or queries that require a web search will fall under this category.
|
56 |
+
If the prompt is phrased with 'What', 'Where', 'How', 'Why',, 'should' followed by phrases that could link back to the dataframe's columns: [date_time,temperature_c,temperature_f,humidity,soil_moisture,light] fall under this category.
|
57 |
+
|
58 |
+
|
59 |
+
Return the category number and name of the prompt in a JSON format with the following format:
|
60 |
+
{{
|
61 |
+
"category_number": 1,
|
62 |
+
"category_name": "Plant Maintenance"
|
63 |
+
}}
|
64 |
+
|
65 |
+
User: {question}
|
66 |
+
"""
|
67 |
|
68 |
return agent.invoke({"input": full_prompt})['output']
|
main.py
CHANGED
@@ -1,17 +1,16 @@
|
|
1 |
from fastapi import FastAPI
|
2 |
|
3 |
import os
|
4 |
-
import
|
5 |
-
import sys
|
6 |
|
7 |
import pandas as pd
|
8 |
from langchain.document_loaders import DirectoryLoader
|
9 |
from agents import master_agent, plant_agent, eda_agent, rag_agent
|
10 |
|
11 |
-
|
12 |
|
13 |
-
os.environ['OPENAI_API_KEY'] = os.
|
14 |
-
os.environ['SERPAPI_API_KEY'] = os.getenv('
|
15 |
|
16 |
master = master_agent.init_config()
|
17 |
|
@@ -31,12 +30,14 @@ loader = DirectoryLoader("data/txt", glob="*.txt")
|
|
31 |
|
32 |
rag = rag_agent.init_config(loader)
|
33 |
|
34 |
-
app = FastAPI()
|
35 |
-
|
36 |
loader = DirectoryLoader("data/txt", glob="*.txt")
|
37 |
|
38 |
documents = loader.load()
|
39 |
|
|
|
|
|
|
|
|
|
40 |
@app.get("/hello")
|
41 |
def hello():
|
42 |
return {"message": "Hello World"}
|
@@ -45,18 +46,24 @@ def hello():
|
|
45 |
def ask(question: str):
|
46 |
category = eval(master_agent.answer_question(master, question))
|
47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
if category['category_number'] == 1:
|
49 |
-
response = eval(plant_agent.answer_question(plant,
|
50 |
|
51 |
category.update(response)
|
52 |
|
53 |
elif category['category_number'] == 2:
|
54 |
-
response = eda_agent.answer_question(eda,
|
55 |
|
56 |
category['response'] = response
|
57 |
|
58 |
elif category['category_number'] == 3:
|
59 |
-
response = rag_agent.answer_question(rag,
|
60 |
|
61 |
category['response'] = response
|
62 |
|
|
|
1 |
from fastapi import FastAPI
|
2 |
|
3 |
import os
|
4 |
+
from dotenv import load_dotenv
|
|
|
5 |
|
6 |
import pandas as pd
|
7 |
from langchain.document_loaders import DirectoryLoader
|
8 |
from agents import master_agent, plant_agent, eda_agent, rag_agent
|
9 |
|
10 |
+
load_dotenv()
|
11 |
|
12 |
+
os.environ['OPENAI_API_KEY'] = os.environ.get("OPENAI_API_KEY")
|
13 |
+
os.environ['SERPAPI_API_KEY'] = os.getenv('SERPAPI_API_KEY')
|
14 |
|
15 |
master = master_agent.init_config()
|
16 |
|
|
|
30 |
|
31 |
rag = rag_agent.init_config(loader)
|
32 |
|
|
|
|
|
33 |
loader = DirectoryLoader("data/txt", glob="*.txt")
|
34 |
|
35 |
documents = loader.load()
|
36 |
|
37 |
+
print("init rag agent")
|
38 |
+
|
39 |
+
app = FastAPI()
|
40 |
+
|
41 |
@app.get("/hello")
|
42 |
def hello():
|
43 |
return {"message": "Hello World"}
|
|
|
46 |
def ask(question: str):
|
47 |
category = eval(master_agent.answer_question(master, question))
|
48 |
|
49 |
+
temp_question = f"Referring to the Blue Indigo False Plant, {question}"
|
50 |
+
|
51 |
+
print(question)
|
52 |
+
print(temp_question)
|
53 |
+
print(category)
|
54 |
+
|
55 |
if category['category_number'] == 1:
|
56 |
+
response = eval(plant_agent.answer_question(plant, temp_question))
|
57 |
|
58 |
category.update(response)
|
59 |
|
60 |
elif category['category_number'] == 2:
|
61 |
+
response = eda_agent.answer_question(eda, temp_question)
|
62 |
|
63 |
category['response'] = response
|
64 |
|
65 |
elif category['category_number'] == 3:
|
66 |
+
response = rag_agent.answer_question(rag, temp_question)
|
67 |
|
68 |
category['response'] = response
|
69 |
|
requirements.txt
CHANGED
@@ -10,4 +10,9 @@ tiktoken
|
|
10 |
langchain
|
11 |
google-search-results
|
12 |
langchainhub
|
13 |
-
unstructured
|
|
|
|
|
|
|
|
|
|
|
|
10 |
langchain
|
11 |
google-search-results
|
12 |
langchainhub
|
13 |
+
unstructured
|
14 |
+
python-dotenv
|
15 |
+
scikit-learn
|
16 |
+
matplotlib
|
17 |
+
seaborn
|
18 |
+
scipy
|
test.py
ADDED
File without changes
|