Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -62,7 +62,7 @@ def find_exact_match(df1, df2, column_name):
|
|
| 62 |
return matches
|
| 63 |
|
| 64 |
|
| 65 |
-
def find_similar_texts(df1, df2, column_name, threshold=0.
|
| 66 |
# Find rows with similar texts in the specified column, excluding exact matches
|
| 67 |
similar_texts = []
|
| 68 |
exact_matches = []
|
|
@@ -91,9 +91,9 @@ def find_similar_texts(df1, df2, column_name, threshold=0.4):
|
|
| 91 |
max_length = max(len(row1[column_name]), len(row2[column_name]))
|
| 92 |
similarity_score = 1 - (distance / max_length)
|
| 93 |
if similarity_score >= threshold:
|
| 94 |
-
if similarity
|
| 95 |
exact_matches.append((i, j, row1[column_name], row2[column_name]))
|
| 96 |
-
|
| 97 |
similar_texts.append((i, j, row1[column_name], row2[column_name]))
|
| 98 |
|
| 99 |
return similar_texts, exact_matches
|
|
@@ -101,6 +101,7 @@ def find_similar_texts(df1, df2, column_name, threshold=0.4):
|
|
| 101 |
|
| 102 |
|
| 103 |
|
|
|
|
| 104 |
def main():
|
| 105 |
st.title("Item Comparison App")
|
| 106 |
|
|
@@ -122,7 +123,7 @@ def main():
|
|
| 122 |
st.header("Select Columns")
|
| 123 |
warehouse_column = st.selectbox("Choose column from warehouse item stocks:", warehouse_columns)
|
| 124 |
industry_column = st.selectbox("Choose column from industry item stocks:", industry_columns)
|
| 125 |
-
|
| 126 |
# Compare button
|
| 127 |
if st.button("Compare"):
|
| 128 |
# Find exact matches
|
|
|
|
| 62 |
return matches
|
| 63 |
|
| 64 |
|
| 65 |
+
def find_similar_texts(df1, df2, column_name, threshold=0.5):
|
| 66 |
# Find rows with similar texts in the specified column, excluding exact matches
|
| 67 |
similar_texts = []
|
| 68 |
exact_matches = []
|
|
|
|
| 91 |
max_length = max(len(row1[column_name]), len(row2[column_name]))
|
| 92 |
similarity_score = 1 - (distance / max_length)
|
| 93 |
if similarity_score >= threshold:
|
| 94 |
+
if similarity >= threshold and similarity < 1: # Exact match
|
| 95 |
exact_matches.append((i, j, row1[column_name], row2[column_name]))
|
| 96 |
+
elif similarity > threshold: # Similar but not the same
|
| 97 |
similar_texts.append((i, j, row1[column_name], row2[column_name]))
|
| 98 |
|
| 99 |
return similar_texts, exact_matches
|
|
|
|
| 101 |
|
| 102 |
|
| 103 |
|
| 104 |
+
|
| 105 |
def main():
|
| 106 |
st.title("Item Comparison App")
|
| 107 |
|
|
|
|
| 123 |
st.header("Select Columns")
|
| 124 |
warehouse_column = st.selectbox("Choose column from warehouse item stocks:", warehouse_columns)
|
| 125 |
industry_column = st.selectbox("Choose column from industry item stocks:", industry_columns)
|
| 126 |
+
|
| 127 |
# Compare button
|
| 128 |
if st.button("Compare"):
|
| 129 |
# Find exact matches
|