Changed to reading in question with JSON file

app.py

@@ -1,13 +1,14 @@
 import gradio as gr
 import os
 
-# PERSISTENT DATA STORAGE: these are used to upload user responses to a dataset
+# PERSISTENT DATA STORAGE: this code is used to make commits 
 
 import json
 from datetime import datetime
 from pathlib import Path
 from uuid import uuid4 
-from huggingface_hub import CommitScheduler
+from huggingface_hub import CommitScheduler, hf_hub_download, file_exists, HfApi
+from random import shuffle
 
 JSON_DATASET_DIR = Path("json_dataset")
 JSON_DATASET_DIR.mkdir(parents=True, exist_ok=True)
@@ -22,55 +23,85 @@ scheduler = CommitScheduler(
     token = os.getenv("HF_TOKEN")
 )
 
-def save_json(score1, score2):
+# Global variables which interact with loading and unloading
+user_data = {}
+current_response = {}
+current_question = {} # read-only within gradio blocks
+user_id = "no_id"
+qIDs = ["mbe_46", "mbe_132", "mbe_287", "mbe_326", "mbe_334", "mbe_389", "mbe_563", "mbe_614", "mbe_642", "mbe_747", "mbe_779", "mbe_826", "mbe_845", "mbe_1042", "mbe_1134"]
+mode_options = ["e5", "colbert"]
+
+def load_user_data(id):
+    global user_data
+    filename = id.replace('@', '_AT_').replace('.', '_DOT_')
+    if file_exists("ebrowne/test-data", "users/" + filename + ".json"):
+        print("File exists, downloading data.")
+        # If the ID exists, download the file from HuggingFace
+        hf_hub_download(repo_id="ebrowne/test-data", token = os.getenv("HF_TOKEN"), filename="users/" + filename + ".json")
+        # Add their current status to user_data
+    else:
+        # If the ID doesn't exist, create a format for the file and upload it to HuggingFace
+        shuffle(qIDs)
+        modes = []
+        for i in range(len(qIDs)): 
+            temp = mode_options[:]
+            shuffle(temp)
+            modes.append(temp)
+        # This is the format for a user's file on HuggingFace
+        user_data = {
+            "user_id": id, # original in email format, which was passed here
+            "order": qIDs, # randomized order for each user
+            "modes": modes, # randomized order for each user
+            "current": 0, # user starts on first question
+            "responses": [] # formatted as a list of current_responses
+        }
+        # Run the update method to upload the new JSON file to HuggingFace
+        update_huggingface(id)
+
+def update_huggingface(id):
+    global user_data
+    filename = id.replace('@', '_AT_').replace('.', '_DOT_')
+    # Create a local file that will be uploaded to HuggingFace
+    with open(filename + ".json", "w") as f:
+        json.dump(user_data, f)
+    # Upload to hub (overwriting existing files...)
+    api = HfApi()
+    api.upload_file(
+        path_or_fileobj=filename + ".json",
+        path_in_repo="users/" + filename + ".json",
+        repo_id="ebrowne/test-data",
+        repo_type="dataset",
+        token = os.getenv("HF_TOKEN")
+    )
+
+def reset_current_response():
+    global current_response
+    current_response = {
+        "user_id": user_id, 
+        "question_id": "QID",
+        "user_answer": 0, 
+        "e5_scores": [], # list of ten [score, score, score, score]
+        "e5_set": [], # two values
+        "e5_generation": [], # two values
+        "colbert_scores": [], 
+        "colbert_set": [],
+        "colbert_generation": [],
+        "gold_set": [],
+        "gold_generation": []
+    }
+
+# This method is being used to save each set of individual scores (in case the main files have issues, the data should be saved)
+def commit_current_and_reset():
     with scheduler.lock:
         with JSON_DATASET_PATH.open("a") as f:
-            json.dump({"relevance": score1, "novelty": score2, "datetime": datetime.now().isoformat()}, f)
+            json.dump(current_response, f)
             f.write("\n")
-
-# READING EXISTING DATA: this is used to read questions
-
-"""
-from datasets import load_dataset
-qa_data = load_dataset("ebrowne/test-data", data_files = "test.json")
-q = qa_data["train"][0] # loaded question data
-"""
+    reset_current_response()
 
 # VARIABLES: will eventually be loaded with JSON from a dataset 
 
-"""
-question_text = q["prompt"] + " " + q["question"]
-answers_text = [q["a"], q["b"], q["c"], q["d"]]
-"""
-question_text = "An act of Congress provides that \"no federal court shall order the implementation of a public school desegregation plan that would require the transportation of any student to a school other than the school closest or next closest to his place of residence.\" Which of the following is the strongest argument for the constitutionality of the act?"
-answers_text = ["The Fourteenth Amendment authorizes Congress to define governmental conduct which violates the equal protection clause.", "Under Article III, Congress may restrict the jurisdiction of the federal courts.", "Transportation of students is subject to regulation by Congress because commerce is involved.", "Congress provides partial support for public education and is therefore entitled to establish conditions upon the expenditure of federal grants."]
-answer_id = 1
-passage_texts = ["The IDEA also requires states receiving federal funds to educate disabled children in the \"[l]east restrictive environment\" appropriate for each child. 20 U.S.C. § 1412(a)(5). The statute mandates at § 1412(a)(5)(A):", 
-            "Under the current iteration of Arkansas' school choice law-the 2017 Act-a school that claims a conflict with participating in school choice must show that it is subject to an active and enforceable desegregation court order or plan that \"explicitly limits the transfer of students between school districts.\" Ark. Code Ann. § 6-18-1906(a)(2). Accordingly, for Junction City to receive an exemption from participating in school choice under the 2017 Act, it must show that it is subject to a desegregation order that explicitly bars \"inter-district\" student transfers.",
-            "This is not to say that a school may unilaterally reject or revise a child's stay-put IEP-that would defang the stay-put requirement entirely. To the contrary, the Supreme Court has made clear that the IDEA \"strip[s] schools of the unilateral authority they had traditionally employed to exclude disabled students.\" Honig , 484 U.S. at 323, 108 S.Ct. 592 (emphasis omitted); see also Sch. Comm. of the Town of Burlington v. Dep't of Educ. , 471 U.S. 359, 373, 105 S.Ct. 1996, 85 L.Ed.2d 385 (1985). Courts should therefore view deviations from the IEP \"with a critical eye to ensure that motivations other than those compatible with the statute, such as bureaucratic inertia, are not driving the decision.\" John M. , 502 F.3d at 715. But context matters; for example, as a child moves \"from elementary school to middle school or from middle school to high school,\" not every change necessitated by a new educational environment will necessarily violate the IDEA. See id. at 714-15.",
-            "Under § 79-209(2), all schools are required to have a policy that states the number of absences after which the school shall render services to address a student's barriers to attendance. Such services shall include, but not be limited to:",
-            "To assure that children attending public schools obtain a high quality education, the legislature shall make adequate provision to ensure that ... there are a sufficient number of classrooms ....",
-            "courts should not disturb a state's denial of IDEA reimbursement where the chief benefits of the chosen school are the kind of advantages that might be preferred by parents of any child, disabled or not. Rather, the unilateral private placement is only appropriate if it provides education instruction [specially ] designed to meet the unique needs of a handicapped child.",
-            "It is also likely that residents of the States will attend out-of-state schools that invoke the Exemptions, and that such students will seek contraceptive services through programs in their home states, also giving rise to fiscal injuries to the States that only a nationwide injunction can remedy.",
-            "Although state classifications based on alienage are generally suspect, a state may reserve a government position for citizens if it is related to self-governance, involves policymaking, or requires exercise of important discretionary power over citizens. In these cases, only a rationality test is used. A public school teacher at the primary and secondary school level performs an important governmental function (e.g., he influences students' attitudes about government, the political process, citizenship, etc.), and therefore the exclusion of aliens is rationally related to the state's interest in furthering educational goals. [Ambach v. Norwick (1979)]",
-            "Passage 9",
-            "Passage 10"]
-generation = "The strongest argument for the constitutionality of the act is that it represents a reasonable balance between achieving school desegregation and respecting local control. It limits federal court orders to prevent overly burdensome transportation requirements for students, aiming to achieve desegregation while minimizing disruption for families. The act is seen as a legitimate exercise of Congress's authority to shape the implementation of federal laws, acknowledging local concerns and practicalities without entirely undermining the desegregation goals."
-gold_passage = "Unlike its original jurisdiction, the appellate jurisdiction of the Supreme Court is subject to “exceptions and regulations” prescribed by Congress, and the jurisdiction of the inferior federal courts is subject to congressional prescription. Additionally, Congress has power to regulate modes and practices of proceeding on the part of the inferior federal courts. Whether there are limitations to the exercise of these congressional powers, and what the limitations may be, are matters that have vexed scholarly and judicial interpretation over the years, inasmuch as congressional displeasure with judicial decisions has sometimes led to successful efforts to “curb” the courts and more frequently to proposed but unsuccessful curbs."
-gold_passage_generation = "The strongest argument for the constitutionality of the Congressional act is that Congress has the constitutional authority to regulate federal court jurisdiction and procedures. This includes setting limits on judicial orders. By restricting desegregation plans to only require transportation to the nearest or next nearest school, Congress is exercising its power to define and control how federal courts address such issues, ensuring judicial actions align with legislative priorities and practical concerns."
-step = 0
-mode = 0
-
-sample_dict = {
-    "question": question_text,
-    "answers": answers_text,
-    "correct_answer_index": answer_id, 
-    "top10": passage_texts,
-    "generation": generation,
-    "top10_contains_gold_passage": False, 
-    "gold_passage": gold_passage,
-    "gold_passage_generation": gold_passage_generation
-}
+with open("example.json", "r") as f:
+    current_question = json.load(f)
 
 # THEMING: colors and styles (Gradio native)
 
@@ -83,8 +114,6 @@ theme = gr.themes.Soft(
 
 # BLOCKS: main user interface
 
-user_id = "NO_ID"
-
 with gr.Blocks(theme = theme) as user_eval:
     # Title text introducing study
     forward_btn = gr.Textbox("unchanged", visible = False, elem_id = "togglebutton") # used for toggling windows
@@ -100,14 +129,14 @@ with gr.Blocks(theme = theme) as user_eval:
         with gr.Column(scale = 2) as passages:
             selection = gr.HTML("""
             <h2> Retrieved Passage </h2>
-            <p> """ + passage_texts[0] + "</p>")
+            <p> """ + current_question["top10_" + user_data["modes"][user_data["current"]][mode]][0] + "</p>")
             line = gr.Markdown("---")
             # New answers is able to render the Q and A with formatting. It doesn't change the contents of the answers.
-            new_answers = answers_text.copy()
-            new_answers[answer_id] = "**" + answers_text[answer_id] + "** ✅"
+            new_answers = current_question["answers"].copy()
+            new_answers[current_question["correct_answer_index"]] = "**" + current_question["answers"][current_question["correct_answer_index"]] + "** ✅"
             passage_display = gr.Markdown("""
             ## Question and Answer
-            *""" + question_text + 
+            *""" + current_question["question"] + 
             """* \n
             + """ + new_answers[0] + 
             """ \n
@@ -150,10 +179,10 @@ with gr.Blocks(theme = theme) as user_eval:
             step += 1
             print(e0)
             print(e1 + e2 + e3)
-            if step == len(passage_texts):
+            if step == len(current_question["top10_" + user_data["modes"][user_data["current"]][mode]]): # should always be 10
                 # Step 10: all sources 
                 collapsible_string = ""
-                for i, passage in enumerate(passage_texts):
+                for i, passage in enumerate(current_question["top10_" + user_data["modes"][user_data["current"]][mode]]):
                     collapsible_string += """
                             <strong>Passage """ + str(i + 1) + """</strong>
                             <p> """ + passage + """ </p>
@@ -171,7 +200,7 @@ with gr.Blocks(theme = theme) as user_eval:
                 return {
                     selection: gr.HTML("""
                         <h2> Retrieved Passage </h2> 
-                        <p> """ + passage_texts[step] + "</p>"),
+                        <p> """ + current_question["top10_" + user_data["modes"][user_data["current"]][mode]][step] + "</p>"),
                     eval_0: gr.Radio(value = None),
                     eval_1: gr.Slider(value = 3),
                     eval_2: gr.Slider(value = 3),
@@ -188,12 +217,12 @@ with gr.Blocks(theme = theme) as user_eval:
                 return {
                     selection: gr.HTML("""
                         <h2> Autogenerated Response </h2>
-                        <p> """ + generation + "</p>"),
+                        <p> """ + current_question["generation_" + user_data["modes"][user_data["current"]][mode]] + "</p>"),
                     eval_helps: gr.Slider(value = 1),
                     eval_satisfied: gr.Slider(value = 1)
                 }
             # Steps 12 and 13 are gold passage + gold passage generation IF it is applicable
-            if step > 11 and not sample_dict["top10_contains_gold_passage"]:
+            if step > 11 and not current_question["top10_contains_gold_passage"]:
                 # When mode is 0 -> reset with mode = 1
                 if mode == 0:
                     return {
@@ -207,7 +236,7 @@ with gr.Blocks(theme = theme) as user_eval:
                     return {
                         selection: gr.HTML("""
                             <h2> Retrieved Passage </h2> 
-                            <p> """ + gold_passage + "</p>"),
+                            <p> """ + current_question["gold_passage"] + "</p>"),
                         forward_btn: gr.Textbox(),
                         eval_helps: gr.Slider(value = 1),
                         eval_satisfied: gr.Slider(value = 1)
@@ -216,7 +245,7 @@ with gr.Blocks(theme = theme) as user_eval:
                    return {
                         selection: gr.HTML("""
                             <h2> Autogenerated Response </h2> 
-                            <p> """ + gold_passage_generation + "</p>"),
+                            <p> """ + current_question["gold_passage_generation"] + "</p>"),
                         forward_btn: gr.Textbox(),
                         eval_helps: gr.Slider(value = 1),
                         eval_satisfied: gr.Slider(value = 1)
@@ -251,11 +280,11 @@ with gr.Blocks(theme = theme) as user_eval:
     with gr.Row(equal_height = False, visible = False) as question:
         with gr.Column():
             gr.Markdown("**Question**")
-            gr.Markdown(question_text)
-            a = gr.Button(answers_text[0])
-            b = gr.Button(answers_text[1])
-            c = gr.Button(answers_text[2])
-            d = gr.Button(answers_text[3])
+            gr.Markdown(current_question["question"])
+            a = gr.Button(current_question["answers"][0])
+            b = gr.Button(current_question["answers"][1])
+            c = gr.Button(current_question["answers"][2])
+            d = gr.Button(current_question["answers"][3])
 
             def answer():
                 return {
@@ -310,7 +339,7 @@ with gr.Blocks(theme = theme) as user_eval:
             def submit_email(email):
                 global user_id
                 user_id = email
-                print(user_id)
+                load_user_data(user_id) # calls login, downloads data, initializes session
                 return {
                     question: gr.Row(visible = True),
                     login: gr.Row(visible = False)

example.json

@@ -0,0 +1,30 @@
+{
+    "question": "An act of Congress provides that \"no federal court shall order the implementation of a public school desegregation plan that would require the transportation of any student to a school other than the school closest or next closest to his place of residence.\" Which of the following is the strongest argument for the constitutionality of the act?",
+    "answers": ["The Fourteenth Amendment authorizes Congress to define governmental conduct which violates the equal protection clause.", "Under Article III, Congress may restrict the jurisdiction of the federal courts.", "Transportation of students is subject to regulation by Congress because commerce is involved.", "Congress provides partial support for public education and is therefore entitled to establish conditions upon the expenditure of federal grants."],
+    "correct_answer_index": 1, 
+    "top10_e5": ["E5 PASSAGE: The IDEA also requires states receiving federal funds to educate disabled children in the \"[l]east restrictive environment\" appropriate for each child. 20 U.S.C. § 1412(a)(5). The statute mandates at § 1412(a)(5)(A):", 
+            "Under the current iteration of Arkansas' school choice law-the 2017 Act-a school that claims a conflict with participating in school choice must show that it is subject to an active and enforceable desegregation court order or plan that \"explicitly limits the transfer of students between school districts.\" Ark. Code Ann. § 6-18-1906(a)(2). Accordingly, for Junction City to receive an exemption from participating in school choice under the 2017 Act, it must show that it is subject to a desegregation order that explicitly bars \"inter-district\" student transfers.",
+            "This is not to say that a school may unilaterally reject or revise a child's stay-put IEP-that would defang the stay-put requirement entirely. To the contrary, the Supreme Court has made clear that the IDEA \"strip[s] schools of the unilateral authority they had traditionally employed to exclude disabled students.\" Honig , 484 U.S. at 323, 108 S.Ct. 592 (emphasis omitted); see also Sch. Comm. of the Town of Burlington v. Dep't of Educ. , 471 U.S. 359, 373, 105 S.Ct. 1996, 85 L.Ed.2d 385 (1985). Courts should therefore view deviations from the IEP \"with a critical eye to ensure that motivations other than those compatible with the statute, such as bureaucratic inertia, are not driving the decision.\" John M. , 502 F.3d at 715. But context matters; for example, as a child moves \"from elementary school to middle school or from middle school to high school,\" not every change necessitated by a new educational environment will necessarily violate the IDEA. See id. at 714-15.",
+            "Under § 79-209(2), all schools are required to have a policy that states the number of absences after which the school shall render services to address a student's barriers to attendance. Such services shall include, but not be limited to:",
+            "To assure that children attending public schools obtain a high quality education, the legislature shall make adequate provision to ensure that ... there are a sufficient number of classrooms ....",
+            "courts should not disturb a state's denial of IDEA reimbursement where the chief benefits of the chosen school are the kind of advantages that might be preferred by parents of any child, disabled or not. Rather, the unilateral private placement is only appropriate if it provides education instruction [specially ] designed to meet the unique needs of a handicapped child.",
+            "It is also likely that residents of the States will attend out-of-state schools that invoke the Exemptions, and that such students will seek contraceptive services through programs in their home states, also giving rise to fiscal injuries to the States that only a nationwide injunction can remedy.",
+            "Although state classifications based on alienage are generally suspect, a state may reserve a government position for citizens if it is related to self-governance, involves policymaking, or requires exercise of important discretionary power over citizens. In these cases, only a rationality test is used. A public school teacher at the primary and secondary school level performs an important governmental function (e.g., he influences students' attitudes about government, the political process, citizenship, etc.), and therefore the exclusion of aliens is rationally related to the state's interest in furthering educational goals. [Ambach v. Norwick (1979)]",
+            "The IEP must aim to enable the child to make progress. After all, the essential function of an IEP is to set out a plan for pursuing academic and functional advancement. See §§ 1414(d)(1)(A)(i)(I)-(IV). This reflects the broad purpose of the IDEA, an \"ambitious\" piece of legislation enacted \"in response to Congress' perception that a majority of handicapped children in the United States 'were either totally excluded from schools or [were] sitting idly in regular classrooms awaiting the time when they were old enough to drop out.' \" Rowley, 458 U.S. at 179, 102 S.Ct. 3034 (quoting H.R. Rep. No. 94-332, p. 2 (1975) ). A substantive standard not focused on student progress would do little to remedy the pervasive and tragic academic stagnation that prompted Congress to act.",
+            "Court to \"exclude\" religious schools from the scholarship program. The provision mandated only that the state treasury not be used to fund religious schooling. As this case demonstrates, that mandate does not necessarily require differential treatment. The no-aid provision can be implemented in two ways. A State may distinguish within a benefit program between secular and sectarian schools, or it may decline to fund all private schools. The Court agrees that the First Amendment permits the latter course. See ante, at 2261 - 2262. Because that is the path the Montana Supreme Court took in this case, there was no reason for this Court to address the alternative."],
+    "generation_e5": "This generation is based on the E5 passages.",
+    "top10_colbert": ["COLBERT PASSAGE: The IDEA also requires states receiving federal funds to educate disabled children in the \"[l]east restrictive environment\" appropriate for each child. 20 U.S.C. § 1412(a)(5). The statute mandates at § 1412(a)(5)(A):", 
+            "Under the current iteration of Arkansas' school choice law-the 2017 Act-a school that claims a conflict with participating in school choice must show that it is subject to an active and enforceable desegregation court order or plan that \"explicitly limits the transfer of students between school districts.\" Ark. Code Ann. § 6-18-1906(a)(2). Accordingly, for Junction City to receive an exemption from participating in school choice under the 2017 Act, it must show that it is subject to a desegregation order that explicitly bars \"inter-district\" student transfers.",
+            "This is not to say that a school may unilaterally reject or revise a child's stay-put IEP-that would defang the stay-put requirement entirely. To the contrary, the Supreme Court has made clear that the IDEA \"strip[s] schools of the unilateral authority they had traditionally employed to exclude disabled students.\" Honig , 484 U.S. at 323, 108 S.Ct. 592 (emphasis omitted); see also Sch. Comm. of the Town of Burlington v. Dep't of Educ. , 471 U.S. 359, 373, 105 S.Ct. 1996, 85 L.Ed.2d 385 (1985). Courts should therefore view deviations from the IEP \"with a critical eye to ensure that motivations other than those compatible with the statute, such as bureaucratic inertia, are not driving the decision.\" John M. , 502 F.3d at 715. But context matters; for example, as a child moves \"from elementary school to middle school or from middle school to high school,\" not every change necessitated by a new educational environment will necessarily violate the IDEA. See id. at 714-15.",
+            "Under § 79-209(2), all schools are required to have a policy that states the number of absences after which the school shall render services to address a student's barriers to attendance. Such services shall include, but not be limited to:",
+            "To assure that children attending public schools obtain a high quality education, the legislature shall make adequate provision to ensure that ... there are a sufficient number of classrooms ....",
+            "courts should not disturb a state's denial of IDEA reimbursement where the chief benefits of the chosen school are the kind of advantages that might be preferred by parents of any child, disabled or not. Rather, the unilateral private placement is only appropriate if it provides education instruction [specially ] designed to meet the unique needs of a handicapped child.",
+            "It is also likely that residents of the States will attend out-of-state schools that invoke the Exemptions, and that such students will seek contraceptive services through programs in their home states, also giving rise to fiscal injuries to the States that only a nationwide injunction can remedy.",
+            "Although state classifications based on alienage are generally suspect, a state may reserve a government position for citizens if it is related to self-governance, involves policymaking, or requires exercise of important discretionary power over citizens. In these cases, only a rationality test is used. A public school teacher at the primary and secondary school level performs an important governmental function (e.g., he influences students' attitudes about government, the political process, citizenship, etc.), and therefore the exclusion of aliens is rationally related to the state's interest in furthering educational goals. [Ambach v. Norwick (1979)]",
+            "The IEP must aim to enable the child to make progress. After all, the essential function of an IEP is to set out a plan for pursuing academic and functional advancement. See §§ 1414(d)(1)(A)(i)(I)-(IV). This reflects the broad purpose of the IDEA, an \"ambitious\" piece of legislation enacted \"in response to Congress' perception that a majority of handicapped children in the United States 'were either totally excluded from schools or [were] sitting idly in regular classrooms awaiting the time when they were old enough to drop out.' \" Rowley, 458 U.S. at 179, 102 S.Ct. 3034 (quoting H.R. Rep. No. 94-332, p. 2 (1975) ). A substantive standard not focused on student progress would do little to remedy the pervasive and tragic academic stagnation that prompted Congress to act.",
+            "Court to \"exclude\" religious schools from the scholarship program. The provision mandated only that the state treasury not be used to fund religious schooling. As this case demonstrates, that mandate does not necessarily require differential treatment. The no-aid provision can be implemented in two ways. A State may distinguish within a benefit program between secular and sectarian schools, or it may decline to fund all private schools. The Court agrees that the First Amendment permits the latter course. See ante, at 2261 - 2262. Because that is the path the Montana Supreme Court took in this case, there was no reason for this Court to address the alternative."],
+    "generation_colbert": "This generation is based on the ColBERT passages.",
+    "top10_contains_gold_passage": false, 
+    "gold_passage": "This is the gold passage.",
+    "gold_passage_generation": "This generation answers the question with the gold passage."
+}