flagging refactor

app.py

@@ -19,31 +19,39 @@ roots_datasets = {
 def get_docid_html(docid):
     data_org, dataset, docid = docid.split("/")
     metadata = roots_datasets[dataset]
+    locked_color = "LightGray"
+    open_color = "#7978FF"
     if metadata.private:
         docid_html = """
         <a title="This dataset is private. See the introductory text for more information"
-            style="color:#AA4A44; font-weight: bold; text-decoration:none"
-            onmouseover="style='color:#AA4A44; font-weight: bold; text-decoration:underline'"
-            onmouseout="style='color:#AA4A44; font-weight: bold; text-decoration:none'"
+            style="color:{locked_color}; font-weight: bold; text-decoration:none"
+            onmouseover="style='color:{locked_color}; font-weight: bold; text-decoration:underline'"
+            onmouseout="style='color:{locked_color}; font-weight: bold; text-decoration:none'"
             href="https://huggingface.co/datasets/bigscience-data/{dataset}"
             target="_blank">
             🔒{dataset}
         </a>
-        <span style="color:#7978FF; ">/{docid}</span>""".format(
-            dataset=dataset, docid=docid
+        <span style="color:{open_color}; ">/{docid}</span>""".format(
+            dataset=dataset,
+            docid=docid,
+            locked_color=locked_color,
+            open_color=open_color,
         )
     else:
         docid_html = """
         <a title="This dataset is licensed {metadata}"
-            style="color:#7978FF; font-weight: bold; text-decoration:none"
-            onmouseover="style='color:#7978FF; font-weight: bold; text-decoration:underline'"
-            onmouseout="style='color:#7978FF; font-weight: bold; text-decoration:none'"
+            style="color:{open_color}; font-weight: bold; text-decoration:none"
+            onmouseover="style='color:{open_color}; font-weight: bold; text-decoration:underline'"
+            onmouseout="style='color:{open_color}; font-weight: bold; text-decoration:none'"
             href="https://huggingface.co/datasets/bigscience-data/{dataset}"
             target="_blank">
             {dataset}
         </a>
-        <span style="color:#7978FF; ">/{docid}</span>""".format(
-            metadata=metadata.tags[0].split(":")[-1], dataset=dataset, docid=docid
+        <span style="color:{open_color}; ">/{docid}</span>""".format(
+            metadata=metadata.tags[0].split(":")[-1],
+            dataset=dataset,
+            docid=docid,
+            open_color=open_color,
         )
     return docid_html
 
@@ -63,30 +71,6 @@ def process_pii(text):
     return text
 
 
-def flag(query, language, num_results, issue_description):
-    try:
-        post_data = {
-            "query": query,
-            "k": num_results,
-            "flag": True,
-            "description": issue_description,
-        }
-        if language != "detect_language":
-            post_data["lang"] = language
-
-        output = requests.post(
-            os.environ.get("address"),
-            headers={"Content-type": "application/json"},
-            data=json.dumps(post_data),
-            timeout=120,
-        )
-
-        results = json.loads(output.text)
-    except:
-        print("Error flagging")
-    return ""
-
-
 def format_result(result, highlight_terms, exact_search, datasets_filter=None):
     text, url, docid = result
     if datasets_filter is not None:
@@ -133,8 +117,9 @@ def format_result(result, highlight_terms, exact_search, datasets_filter=None):
     language = "FIXME"
     result_html = """{}
         <span style='font-size:14px; font-family: Arial; color:#7978FF; text-align: left;'>Document ID: {}</span>
-        <a href="https://forms.gle/AdBLLwRApqcLkHYA8" target="_blank"><button>🏴‍☠️ Flag result 🏴‍☠️</button></a><br>
-
+        <a href="https://forms.gle/AdBLLwRApqcLkHYA8" target="_blank">
+            <button style="color:#ffcdf8; ">🏴‍☠️ Flag result 🏴‍☠️</button>
+        </a><br>
         <!--  <span style='font-size:12px; font-family: Arial; color:MediumAquaMarine'>Language: {}</span><br> -->
         <span style='font-family: Arial;'>{}</span><br>
         <br>
@@ -273,9 +258,7 @@ information and instructions on how to access the full corpus check [this form](
 
 
 if __name__ == "__main__":
-    demo = gr.Blocks(
-        css=".underline-on-hover:hover { text-decoration: underline; } .flagging { font-size:12px; color:Silver; }"
-    )
+    demo = gr.Blocks(css=".underline-on-hover:hover { text-decoration: underline; }")
 
     with demo:
         processed_results_state = gr.State([])
@@ -283,7 +266,6 @@ if __name__ == "__main__":
         num_results_state = gr.State(0)
         exact_search_state = gr.State(False)
         lang_state = gr.State("")
-        max_page_size_state = gr.State(100)
         received_results_state = gr.State(0)
 
         with gr.Row():
@@ -319,7 +301,13 @@ if __name__ == "__main__":
                 value="en",
                 label="Language",
             )
-            k = gr.Slider(1, 100, value=10, step=1, label="Max Results")
+            k = gr.Slider(
+                1,
+                100,
+                value=10,
+                step=1,
+                label="Max Results in fuzzy search or Max Results per page in exact search",
+            )
         with gr.Row():
             submit_btn = gr.Button("Submit")
         with gr.Row(visible=False) as datasets_filter:
@@ -336,19 +324,17 @@ if __name__ == "__main__":
         with gr.Row(visible=False) as pagination:
             next_page_btn = gr.Button("Next Page")
 
-        def run_query(query, lang, k, dropdown_input, max_page_size, received_results):
+        def run_query(query, lang, k, dropdown_input, received_results):
             query = query.strip()
             exact_search = False
             if query.startswith('"') and query.endswith('"') and len(query) >= 2:
                 exact_search = True
                 query = query[1:-1]
-                k = max_page_size
             else:
                 query = " ".join(query.split())
             if query == "" or query is None:
                 return None
 
-            print("submitting", query, lang, k)
             payload = request_payload(query, lang, exact_search, k, received_results)
             err = extract_error_from_payload(payload)
             if err is not None:
@@ -377,7 +363,8 @@ if __name__ == "__main__":
                 ds,
             )
 
-        def submit(query, lang, k, dropdown_input, max_page_size):
+        def submit(query, lang, k, dropdown_input):
+            print("submitting", query, lang, k)
             (
                 processed_results,
                 highlight_terms,
@@ -385,8 +372,8 @@ if __name__ == "__main__":
                 exact_search,
                 result_page_html,
                 datasets,
-            ) = run_query(query, lang, k, dropdown_input, max_page_size, 0)
-            has_more_results = exact_search and (num_results > max_page_size)
+            ) = run_query(query, lang, k, dropdown_input, 0)
+            has_more_results = exact_search and (num_results > k)
             return [
                 processed_results,
                 highlight_terms,
@@ -404,7 +391,6 @@ if __name__ == "__main__":
             lang,
             k,
             dropdown_input,
-            max_page_size,
             received_results,
             processed_results,
         ):
@@ -415,11 +401,9 @@ if __name__ == "__main__":
                 exact_search,
                 result_page_html,
                 datasets,
-            ) = run_query(
-                query, lang, k, dropdown_input, max_page_size, received_results
-            )
+            ) = run_query(query, lang, k, dropdown_input, received_results)
             num_processed_results = len(next(iter(processed_results.values())))
-            has_more_results = exact_search and (num_results > max_page_size)
+            has_more_results = exact_search and (num_results > k)
             print("num_processed_results", num_processed_results)
             print("has_more_results", has_more_results)
             print("received_results", received_results)
@@ -430,9 +414,7 @@ if __name__ == "__main__":
                 exact_search,
                 gr.update(visible=True),
                 gr.Dropdown.update(choices=datasets, value=datasets),
-                gr.update(
-                    visible=num_processed_results >= max_page_size and has_more_results
-                ),
+                gr.update(visible=num_processed_results >= k and has_more_results),
                 received_results + num_processed_results,
                 result_page_html,
             ]
@@ -457,7 +439,7 @@ if __name__ == "__main__":
 
         query.submit(
             fn=submit,
-            inputs=[query, lang, k, available_datasets, max_page_size_state],
+            inputs=[query, lang, k, available_datasets],
             outputs=[
                 processed_results_state,
                 highlight_terms_state,
@@ -472,7 +454,7 @@ if __name__ == "__main__":
         )
         submit_btn.click(
             submit,
-            inputs=[query, lang, k, available_datasets, max_page_size_state],
+            inputs=[query, lang, k, available_datasets],
             outputs=[
                 processed_results_state,
                 highlight_terms_state,
@@ -493,7 +475,6 @@ if __name__ == "__main__":
                 lang,
                 k,
                 available_datasets,
-                max_page_size_state,
                 received_results_state,
                 processed_results_state,
             ],