Update app.py
Browse files
app.py
CHANGED
@@ -3,8 +3,14 @@ from PIL import Image
|
|
3 |
import pytesseract
|
4 |
import pandas as pd
|
5 |
import re
|
|
|
6 |
|
7 |
|
|
|
|
|
|
|
|
|
|
|
8 |
def extract_text(image):
|
9 |
"""
|
10 |
Extract text from the image using Tesseract.
|
@@ -12,18 +18,16 @@ def extract_text(image):
|
|
12 |
return pytesseract.image_to_string(image)
|
13 |
|
14 |
|
|
|
15 |
def clean_and_parse_extracted_text(raw_text):
|
16 |
"""
|
17 |
Parse and clean the raw text to extract structured data.
|
18 |
"""
|
19 |
-
# Split the text into lines and clean up
|
20 |
lines = raw_text.split("\n")
|
21 |
lines = [line.strip() for line in lines if line.strip()]
|
22 |
|
23 |
-
# Identify and extract rows with valid components
|
24 |
data = []
|
25 |
for line in lines:
|
26 |
-
# Match rows containing numeric ranges and values
|
27 |
match = re.match(
|
28 |
r"^(.*?)(\d+(\.\d+)?)(\s*-?\s*\d+(\.\d+)?\s*-?\s*\d+(\.\d+)?)?\s*([a-zA-Z/%]+)?\s*(H|L|Normal)?$",
|
29 |
line,
|
@@ -42,21 +46,16 @@ def clean_and_parse_extracted_text(raw_text):
|
|
42 |
unit = match.group(7)
|
43 |
flag = "Normal" # Default flag
|
44 |
|
45 |
-
# Determine the flag based on value and range
|
46 |
if min_val is not None and max_val is not None:
|
47 |
if value < min_val:
|
48 |
flag = "L"
|
49 |
elif value > max_val:
|
50 |
flag = "H"
|
51 |
|
52 |
-
# Only append the data if the flag is abnormal (L or H)
|
53 |
if flag != "Normal":
|
54 |
data.append([component, value, min_val, max_val, unit, flag])
|
55 |
|
56 |
-
# Create a DataFrame
|
57 |
df = pd.DataFrame(data, columns=["Component", "Your Value", "Min", "Max", "Units", "Flag"])
|
58 |
-
|
59 |
-
# Fix misspellings and inconsistencies (if any known issues exist)
|
60 |
correction_map = {
|
61 |
"emoglobin": "Hemoglobin",
|
62 |
"ematocrit": "Hematocrit",
|
@@ -69,16 +68,43 @@ def clean_and_parse_extracted_text(raw_text):
|
|
69 |
return df
|
70 |
|
71 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
72 |
def display_results(df):
|
73 |
"""
|
74 |
-
Display the flagged abnormalities in a table format.
|
75 |
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
76 |
st.dataframe(df, use_container_width=True)
|
|
|
|
|
|
|
77 |
|
78 |
|
79 |
# Streamlit app
|
80 |
-
st.title("Blood Report Analyzer")
|
81 |
-
st.write("Upload an image of a blood test report to analyze.")
|
82 |
|
83 |
uploaded_file = st.file_uploader("Upload Image", type=["png", "jpg", "jpeg"])
|
84 |
|
@@ -96,8 +122,7 @@ if uploaded_file is not None:
|
|
96 |
# Parse the extracted text into a structured format
|
97 |
parsed_data = clean_and_parse_extracted_text(extracted_text)
|
98 |
|
99 |
-
# Display the structured data
|
100 |
-
st.subheader("Flagged Abnormalities")
|
101 |
display_results(parsed_data)
|
102 |
|
103 |
except Exception as e:
|
|
|
3 |
import pytesseract
|
4 |
import pandas as pd
|
5 |
import re
|
6 |
+
import openai
|
7 |
|
8 |
|
9 |
+
# Setup OpenAI API key (replace with your OpenAI API key)
|
10 |
+
openai.api_key = "your-api-key-here"
|
11 |
+
|
12 |
+
|
13 |
+
# Function to extract text from the image
|
14 |
def extract_text(image):
|
15 |
"""
|
16 |
Extract text from the image using Tesseract.
|
|
|
18 |
return pytesseract.image_to_string(image)
|
19 |
|
20 |
|
21 |
+
# Function to parse and clean the extracted text
|
22 |
def clean_and_parse_extracted_text(raw_text):
|
23 |
"""
|
24 |
Parse and clean the raw text to extract structured data.
|
25 |
"""
|
|
|
26 |
lines = raw_text.split("\n")
|
27 |
lines = [line.strip() for line in lines if line.strip()]
|
28 |
|
|
|
29 |
data = []
|
30 |
for line in lines:
|
|
|
31 |
match = re.match(
|
32 |
r"^(.*?)(\d+(\.\d+)?)(\s*-?\s*\d+(\.\d+)?\s*-?\s*\d+(\.\d+)?)?\s*([a-zA-Z/%]+)?\s*(H|L|Normal)?$",
|
33 |
line,
|
|
|
46 |
unit = match.group(7)
|
47 |
flag = "Normal" # Default flag
|
48 |
|
|
|
49 |
if min_val is not None and max_val is not None:
|
50 |
if value < min_val:
|
51 |
flag = "L"
|
52 |
elif value > max_val:
|
53 |
flag = "H"
|
54 |
|
|
|
55 |
if flag != "Normal":
|
56 |
data.append([component, value, min_val, max_val, unit, flag])
|
57 |
|
|
|
58 |
df = pd.DataFrame(data, columns=["Component", "Your Value", "Min", "Max", "Units", "Flag"])
|
|
|
|
|
59 |
correction_map = {
|
60 |
"emoglobin": "Hemoglobin",
|
61 |
"ematocrit": "Hematocrit",
|
|
|
68 |
return df
|
69 |
|
70 |
|
71 |
+
# Function to generate AI-powered recommendations using OpenAI GPT
|
72 |
+
def generate_medical_recommendation(test_results):
|
73 |
+
"""
|
74 |
+
Generate medical recommendations using OpenAI GPT model based on abnormal test results.
|
75 |
+
"""
|
76 |
+
# Create a structured input for the model
|
77 |
+
prompt = f"Given the following blood test results: {test_results}, provide medical recommendations for a patient."
|
78 |
+
|
79 |
+
response = openai.Completion.create(
|
80 |
+
model="gpt-4", # Use GPT-4 for medical-based responses
|
81 |
+
prompt=prompt,
|
82 |
+
max_tokens=150
|
83 |
+
)
|
84 |
+
|
85 |
+
return response.choices[0].text.strip()
|
86 |
+
|
87 |
+
|
88 |
+
# Function to display results
|
89 |
def display_results(df):
|
90 |
"""
|
91 |
+
Display the flagged abnormalities and medical recommendations in a table format.
|
92 |
"""
|
93 |
+
# Generate a summary of the abnormal test results for the AI model
|
94 |
+
abnormal_results = df[df['Flag'] != 'Normal'].to_string(index=False)
|
95 |
+
recommendation = generate_medical_recommendation(abnormal_results)
|
96 |
+
|
97 |
+
# Display the DataFrame and the recommendation
|
98 |
+
st.subheader("Flagged Abnormalities")
|
99 |
st.dataframe(df, use_container_width=True)
|
100 |
+
|
101 |
+
st.subheader("Medical Recommendations from AI")
|
102 |
+
st.write(recommendation)
|
103 |
|
104 |
|
105 |
# Streamlit app
|
106 |
+
st.title("Blood Report Analyzer with AI Recommendations")
|
107 |
+
st.write("Upload an image of a blood test report to analyze and get AI-powered recommendations.")
|
108 |
|
109 |
uploaded_file = st.file_uploader("Upload Image", type=["png", "jpg", "jpeg"])
|
110 |
|
|
|
122 |
# Parse the extracted text into a structured format
|
123 |
parsed_data = clean_and_parse_extracted_text(extracted_text)
|
124 |
|
125 |
+
# Display the structured data with AI recommendations
|
|
|
126 |
display_results(parsed_data)
|
127 |
|
128 |
except Exception as e:
|