fix reference new line

app.py

@@ -101,7 +101,7 @@ def split_text_from_refs(text: str, sep="\n"):
         for i, ref in enumerate(references, 1):
             ref = remove_bracketed_numbers(ref)
             formatted_refs.append(f"[{i}] {ref}{sep}")
-        formatted_refs = f"{sep}{sep}References:{sep}" + f"{sep}".join(formatted_refs)
+        formatted_refs = f"{sep}{sep}References:{sep}{sep}" + f"{sep}".join(formatted_refs)
     else:
         formatted_refs = ""
 
@@ -240,9 +240,9 @@ def ai_generated_test_gptzero(text):
 
 
 def highlighter_polygraf(text, model="Polygraf AI (Base Model)"):
-    body, references = split_text_from_refs(text, "<br>")
+    body, references = split_text_from_refs(text)
     score, text = detection_polygraf(text=body, model=model)
-    text = text + references
+    text = text + references.replace("\n", "<br>")
     return score, text
 
 

locustfile.py

@@ -0,0 +1,126 @@
+from locust import HttpUser, task, between
+from gradio_client import Client, handle_file
+
+
+class GradioAppUser(HttpUser):
+    wait_time = between(1, 3)
+
+    def on_start(self):
+        self.client = Client("http://sd.demo.polygraf.ai:7890/")
+
+    @task
+    def test_generate_text_with_google_search_and_pdf(self):
+        generated_text = self.client.predict(
+            input_role="Student",
+            topic="Low Resource Large Language Models",
+            keywords="",
+            article_length=350,
+            format="Article",
+            writing_style="Formal",
+            tone="Professional",
+            user_category="General Public",
+            depth_of_content="Moderate analysis",
+            structure="Introduction, Body, Conclusion",
+            references="News outlets",
+            num_examples="1-2",
+            conclusion_type="Call to Action",
+            # ai_model="LLaMA 3",
+            ai_model="OpenAI GPT 4o Mini",
+            google_search_check=False,
+            year_from="2000",
+            month_from="January",
+            day_from="01",
+            year_to="2024",
+            month_to="August",
+            day_to="08",
+            domains_to_include=["com", "org", "net", "int", "edu", "gov", "mil"],
+            include_sites="",
+            exclude_sites="",
+            pdf_file_input=None,
+            api_name="/generate_and_format",
+        )
+
+        self.client.predict(text="Polygraf AI (Advanced Model)", api_name="/highlight_visible_1")
+        detect_generated = self.client.predict(
+            text=generated_text,
+            option="Polygraf AI (Advanced Model)",
+            api_name="/ai_check",
+        )
+
+        humanized_text = self.client.predict(
+            text=generated_text,
+            model="XL Model",
+            temperature=1.2,
+            repetition_penalty=1,
+            top_k=50,
+            length_penalty=1,
+            api_name="/humanize",
+        )
+
+        self.client.predict(text="Polygraf AI (Advanced Model)", api_name="/highlight_visible_1")
+        detect_humanized = self.client.predict(
+            text=humanized_text,
+            option="Polygraf AI (Advanced Model)",
+            api_name="/ai_check",
+        )
+
+        print(detect_generated)
+        print(detect_humanized)
+
+    @task
+    def test_generate_text_without_google_search_and_pdf(self):
+        generated_text = self.client.predict(
+            input_role="Student",
+            topic="Low Resource Large Language Models",
+            keywords="",
+            article_length=400,
+            format="Article",
+            writing_style="Formal",
+            tone="Professional",
+            user_category="General Public",
+            depth_of_content="Moderate analysis",
+            structure="Introduction, Body, Conclusion",
+            references="News outlets",
+            num_examples="1-2",
+            conclusion_type="Call to Action",
+            ai_model="LLaMA 3",
+            google_search_check=False,
+            year_from="2000",
+            month_from="January",
+            day_from="01",
+            year_to="2024",
+            month_to="August",
+            day_to="08",
+            domains_to_include=["com", "org", "net", "int", "edu", "gov", "mil"],
+            include_sites="",
+            exclude_sites="",
+            pdf_file_input=[handle_file("/home/eljan/article_writer/Abstract.pdf")],
+            api_name="/generate_and_format",
+        )
+
+        self.client.predict(text="Polygraf AI (Advanced Model)", api_name="/highlight_visible_1")
+        detect_generated = self.client.predict(
+            text=generated_text,
+            option="Polygraf AI (Advanced Model)",
+            api_name="/ai_check",
+        )
+
+        humanized_text = self.client.predict(
+            text=generated_text,
+            model="XL Model",
+            temperature=1.2,
+            repetition_penalty=1,
+            top_k=50,
+            length_penalty=1,
+            api_name="/humanize",
+        )
+
+        self.client.predict(text="Polygraf AI (Advanced Model)", api_name="/highlight_visible_1")
+        detect_humanized = self.client.predict(
+            text=humanized_text,
+            option="Polygraf AI (Advanced Model)",
+            api_name="/ai_check",
+        )
+
+        print(detect_generated[0]["confidences"])
+        print(detect_humanized[0]["confidences"])

stress_test.py

@@ -0,0 +1,60 @@
+from gradio_client import Client, file, handle_file
+
+client = Client("http://sd.demo.polygraf.ai:7890/")
+
+generated_text = client.predict(
+    input_role="Student",
+    topic="Low Resource Language Models",
+    keywords="",
+    article_length=400,
+    format="Article",
+    writing_style="Formal",
+    tone="Professional",
+    user_category="General Public",
+    depth_of_content="Moderate analysis",
+    structure="Introduction, Body, Conclusion",
+    references="News outlets",
+    num_examples="1-2",
+    conclusion_type="Call to Action",
+    # ai_model="OpenAI GPT 4o Mini",
+    ai_model="LLaMA 3",
+    google_search_check=True,
+    year_from="2000",
+    month_from="January",
+    day_from="01",
+    year_to="2024",
+    month_to="August",
+    day_to="08",
+    domains_to_include=["com", "org", "net", "int", "edu", "gov", "mil"],
+    include_sites="",
+    exclude_sites="",
+    pdf_file_input=[handle_file("/home/eljan/article_writer/Abstract.pdf")],
+    api_name="/generate_and_format",
+)
+
+client.predict(text="Polygraf AI (Advanced Model)", api_name="/highlight_visible_1")
+detect_generated = client.predict(
+    text=generated_text,
+    option="Polygraf AI (Advanced Model)",
+    api_name="/ai_check",
+)
+
+humanized_text = client.predict(
+    text=generated_text,
+    model="XL Model",
+    temperature=1.2,
+    repetition_penalty=1,
+    top_k=50,
+    length_penalty=1,
+    api_name="/humanize",
+)
+
+client.predict(text="Polygraf AI (Advanced Model)", api_name="/highlight_visible_1")
+detect_humanized = client.predict(
+    text=humanized_text,
+    option="Polygraf AI (Advanced Model)",
+    api_name="/ai_check",
+)
+
+print(detect_generated)
+print(detect_humanized)