File size: 4,465 Bytes
b295f13 8a93e99 b295f13 8a93e99 b295f13 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 |
import nltk
nltk.download("all", quiet=True)
from datetime import datetime
import random
import json
from . import preprocess_data
from . import qa_agent
PROMPT = """
Task: Carefully review the given patent text and extract as much physical measurements information such as length/distance, mass/weight, time, temperature, Volume, area, speed, pressure, energy, power, electric current
and voltage, frequency, force, acceleration, density, resistivity, magnetic field strength, and luminous intensity as much as possible.
We are particularly interested in physical measurements including substance that was measured, Value of the measurement, and Unit of the measurement, and measurement type mentioned in the text.
For each measurement, please provide the following details:
- The substance that was measured. (substance)
- The specific value or range that was measured. (Measured Value)
- The unit of the measurement, if provided. (Unit)
- The type of measurement being conducted (e.g., diameter, size, etc.)
Format your response in a structured JSON-like format, as follows:
{"Content": [
{
"Measurement_substance": "substance",
"Measured_value": "value",
"Measured_unit": "unit",
"measurement_type": "type"
},
// ... additional measurements, if present
]
}
If multiple measurements are present in the text, each should be listed as a separate object within the "Content" array.
Example: If the text includes the sentence, "The resulting BaCO3 had a crystallite size of between about 20 and 40 nm", the output should be:
{"Content": [
{
"Measurement_substance": "BaCO3",
"Measured_value": "between about 20 and 40",
"Measured_unit": "nm",
"measurement_type": "crystallite size"
}
]
}
Try to provide as complete and accurate information as possible. Print only the formatted JSON response.
"""
def main():
"""
Main function to:
- Authenticate with OpenAI
- Receive and parse date input from the user
- Extract and print year, month, day
- Preprocess patent data
- Analyze selected patents using GPT-3.5 Turbo
- Print results including cost and optionally output
"""
print("Starting the patent analysis process...")
# Step 1: Input the date from the user
user_date_input = input("Enter a date in the format 'YYYY-MM-DD': ")
# Step 2: Parse the input date into a datetime object
try:
input_date = datetime.strptime(user_date_input, "%Y-%m-%d")
except ValueError:
print(
"Invalid date format. Please enter a valid date in the format 'YYYY-MM-DD'."
)
return
# Step 3: Extract date components
year = input_date.year
month = input_date.month
day = input_date.day
print("Year:", year)
print("Month:", month)
print("Day:", day)
# Step 4: Get random patents number from user
num_patents_to_analyze = int(
input("Enter the number of patents you want to analyze: ")
)
logging_choice = input("Do you want to log the results? (yes/no): ").strip().lower()
logging_enabled = logging_choice == "yes"
model_choice = input(
"Select a model for analysis: 1. gpt-3.5-turbo 2. gpt-4"
).strip()
if model_choice == "1":
model_name = "gpt-3.5-turbo"
elif model_choice == "2":
model_name = "gpt-4"
else:
print("Invalid choice, defaulting to gpt-3.5-turbo.")
model_name = "gpt-3.5-turbo"
print("Processing patents...")
# Step 5: Parse and save patents
saved_patent_names = preprocess_data.parse_and_save_patents(year, month, day, False)
# Step 6: Select random patents and analyze
random_patents = random.sample(saved_patent_names, num_patents_to_analyze)
gpt_3_results = {}
total_cost_gpt3 = 0
# Step 7: Process patents with GPT-3.5 Turbo
for i in range(len(random_patents)):
cost, output = qa_agent.call_QA_to_json(
PROMPT, year, month, day, random_patents, i, logging_enabled, model_name
)
total_cost_gpt3 += cost
average_cost_gpt3 = total_cost_gpt3 / num_patents_to_analyze
print("Patent analysis process completed successfully.")
# Step 8: Print results
print("\nResults for GPT-3.5 Turbo:")
print("Number of patents analyzed:", num_patents_to_analyze)
print("Total cost for analyzing all patents:", total_cost_gpt3)
print("Average cost per patent:", average_cost_gpt3) |