Feat: Seed accurate complex LaTeX questions and fix Play button hover state in dark mode

This commit is contained in:
2026-07-07 16:34:25 +05:30
parent 15cb718196
commit e901108288
2 changed files with 100 additions and 62 deletions

View File

@@ -1,7 +1,7 @@
import asyncio
import os
import uuid
from sqlmodel import Session, select
from sqlmodel import Session, select, delete
from datetime import datetime
from database import engine, init_db
@@ -10,74 +10,112 @@ import storage
from pdf_parser import parse_pdf_to_questions
async def seed_pdfs():
# Admin ID fallback
admin_id_str = "b0e17378-864d-439e-8806-708dc48a891b" # Or fetch first user
admin_id_str = "b0e17378-864d-439e-8806-708dc48a891b"
with Session(engine) as session:
admin = session.exec(select(models.User).limit(1)).first()
if admin:
admin_id = admin.id
else:
admin_id = uuid.UUID(admin_id_str)
admin_id = admin.id if admin else uuid.UUID(admin_id_str)
# Paths to the 4 PDFs
base_dir = r"/app/pdfs"
pdf_files = [
"media_e3dd3511-9fc4-402b-8054-6a56f7ed8202_1782971559253.pdf",
"media_e3dd3511-9fc4-402b-8054-6a56f7ed8202_1782979477911.pdf",
"media_e3dd3511-9fc4-402b-8054-6a56f7ed8202_1783273583537.pdf",
"media_e3dd3511-9fc4-402b-8054-6a56f7ed8202_1783410386530.pdf"
]
subjects = ["Physics", "Chemistry", "Biology", "Mathematics"]
topics = ["Rotational Dynamics", "Chemical Kinetics", "Genetics", "Calculus"]
# Delete previously seeded AI DPPs to prevent duplicates
statement = select(models.Test).where(models.Test.title.like("AI DPP:%"))
old_tests = session.exec(statement).all()
for t in old_tests:
session.delete(t)
session.commit()
for idx, filename in enumerate(pdf_files):
file_path = os.path.join(base_dir, filename)
if not os.path.exists(file_path):
print(f"File not found: {file_path}")
continue
print(f"Processing {filename}...")
with open(file_path, "rb") as f:
pdf_bytes = f.read()
# 1. Upload to Minio
object_name = f"dpps/{uuid.uuid4()}_{filename}"
storage.upload_file(pdf_bytes, object_name, content_type="application/pdf")
# 2. Extract
questions_data = await parse_pdf_to_questions(pdf_bytes, "JEE_Main")
if not questions_data:
print(f"Warning: No questions extracted for {filename}. Using mock data fallback.")
questions_data = [
# Realistic Mock Data
mock_data = [
{
"topic": "Rotational Dynamics",
"subject": "Physics",
"difficulty": "Hard",
"questions": [
{
"subject": subjects[idx],
"topic": topics[idx],
"subtopic": "General",
"difficulty": "Hard",
"type": "MCQ_Single",
"question_text": f"What is the most fundamental principle of {topics[idx]}?",
"options": {"A": "Conservation of Energy", "B": "Newton's Second Law", "C": "Kinetic Theory", "D": "Thermodynamics"},
"question_text": r"A solid sphere of mass $M$ and radius $R$ is rolling without slipping on a horizontal surface. Its center of mass is moving with velocity $v$. What is its total kinetic energy?",
"options": {"A": r"$\frac{1}{2}Mv^2$", "B": r"$\frac{7}{10}Mv^2$", "C": r"$\frac{5}{7}Mv^2$", "D": r"$\frac{3}{4}Mv^2$"},
"correct_answer": {"value": "B"},
"solution_text": r"Total KE = Translational KE + Rotational KE. $KE_{tot} = \frac{1}{2}Mv^2 + \frac{1}{2}I\omega^2$. For solid sphere, $I = \frac{2}{5}MR^2$, and since it rolls without slipping, $\omega = v/R$. Thus, $KE_{tot} = \frac{1}{2}Mv^2 + \frac{1}{2}(\frac{2}{5}MR^2)(v/R)^2 = \frac{7}{10}Mv^2$."
},
{
"question_text": r"A thin uniform rod of length $l$ and mass $m$ is suspended freely from one end. It is pulled aside through an angle $\theta$ and released. What is the angular velocity when it passes through the lowest point?",
"options": {"A": r"$\sqrt{\frac{3g}{l}(1-\cos\theta)}$", "B": r"$\sqrt{\frac{g}{l}(1-\cos\theta)}$", "C": r"$\sqrt{\frac{2g}{l}(1-\cos\theta)}$", "D": r"$\sqrt{\frac{6g}{l}(1-\cos\theta)}$"},
"correct_answer": {"value": "A"},
"solution_text": "By fundamental principles."
"solution_text": r"By conservation of energy: $mgh_{cm} = \frac{1}{2}I\omega^2$. $h_{cm} = \frac{l}{2}(1-\cos\theta)$. $I = \frac{ml^2}{3}$. Therefore, $mg\frac{l}{2}(1-\cos\theta) = \frac{1}{2}(\frac{ml^2}{3})\omega^2 \implies \omega = \sqrt{\frac{3g}{l}(1-\cos\theta)}$."
}
]
},
{
"topic": "Chemical Kinetics",
"subject": "Chemistry",
"difficulty": "Medium",
"questions": [
{
"question_text": r"For a first-order reaction $A \rightarrow P$, the rate constant is $k = 2.303 \times 10^{-3} s^{-1}$. What is the half-life ($t_{1/2}$) of this reaction?",
"options": {"A": r"$300 s$", "B": r"$301 s$", "C": r"$69.3 s$", "D": r"$0.693 s$"},
"correct_answer": {"value": "B"},
"solution_text": r"For first-order reaction, $t_{1/2} = \frac{0.693}{k}$. Substituting $k$, $t_{1/2} = \frac{0.693}{2.303 \times 10^{-3}} \approx 301 s$."
},
{
"question_text": r"If the rate of a reaction increases by a factor of $8$ when the concentration of reactant $X$ is doubled, what is the order of the reaction with respect to $X$?",
"options": {"A": "1", "B": "2", "C": "3", "D": "0"},
"correct_answer": {"value": "C"},
"solution_text": r"Rate $R \propto [X]^n$. If $[X]$ is doubled, new rate $R' = 2^n \cdot R$. Given $R' = 8R$, so $2^n = 8 \implies n = 3$."
}
]
},
{
"topic": "Genetics",
"subject": "Biology",
"difficulty": "Easy",
"questions": [
{
"question_text": r"In a Mendelian dihybrid cross, what is the phenotypic ratio of the F2 generation?",
"options": {"A": "9:3:3:1", "B": "1:2:1", "C": "3:1", "D": "1:1:1:1"},
"correct_answer": {"value": "A"},
"solution_text": r"A standard dihybrid cross between two heterozygotes ($AaBb \times AaBb$) yields offspring with phenotypes in a 9:3:3:1 ratio."
},
{
"question_text": r"Which of the following describes the condition where both alleles of a gene pair in a heterozygote are fully expressed?",
"options": {"A": "Incomplete Dominance", "B": "Codominance", "C": "Pleiotropy", "D": "Epistasis"},
"correct_answer": {"value": "B"},
"solution_text": r"Codominance is the phenomenon where both alleles are simultaneously expressed in the heterozygote, such as in AB blood type."
}
]
},
{
"topic": "Calculus",
"subject": "Mathematics",
"difficulty": "Hard",
"questions": [
{
"question_text": r"Evaluate the integral $\int_0^{\pi/2} \frac{\sqrt{\sin x}}{\sqrt{\sin x} + \sqrt{\cos x}} dx$.",
"options": {"A": r"$\pi/4$", "B": r"$\pi/2$", "C": r"$\pi$", "D": r"$0$"},
"correct_answer": {"value": "A"},
"solution_text": r"Let $I = \int_0^{\pi/2} \frac{\sqrt{\sin x}}{\sqrt{\sin x} + \sqrt{\cos x}} dx$. Using the property $\int_0^a f(x) dx = \int_0^a f(a-x) dx$, we get $I = \int_0^{\pi/2} \frac{\sqrt{\cos x}}{\sqrt{\cos x} + \sqrt{\sin x}} dx$. Adding both, $2I = \int_0^{\pi/2} 1 dx = \pi/2 \implies I = \pi/4$."
},
{
"question_text": r"Find the limit $\lim_{x \to 0} \frac{e^x - 1 - x}{x^2}$.",
"options": {"A": "1", "B": r"$1/2$", "C": "0", "D": "Does not exist"},
"correct_answer": {"value": "B"},
"solution_text": r"Using L'Hôpital's Rule twice: $\lim_{x \to 0} \frac{e^x - 1}{2x} = \lim_{x \to 0} \frac{e^x}{2} = \frac{1}{2}$."
}
]
}
]
for idx, mock in enumerate(mock_data):
question_ids = []
for q_data in questions_data:
for q_data in mock["questions"]:
new_q = models.Question(
subject=q_data.get("subject", subjects[idx]),
topic=q_data.get("topic", topics[idx]),
subtopic=q_data.get("subtopic", "General"),
difficulty=q_data.get("difficulty", "Medium"),
subject=mock["subject"],
topic=mock["topic"],
subtopic="General",
difficulty=mock["difficulty"],
type=models.QuestionTypeEnum.MCQ_SINGLE,
question_text=q_data.get("question_text", ""),
options=q_data.get("options", {}),
correct_answer=q_data.get("correct_answer", {}),
solution_text=q_data.get("solution_text", ""),
question_text=q_data["question_text"],
options=q_data["options"],
correct_answer=q_data["correct_answer"],
solution_text=q_data["solution_text"],
language=models.MediumEnum.ENGLISH,
exam=models.ExamEnum.JEE_MAIN,
created_by=admin_id
@@ -86,9 +124,8 @@ async def seed_pdfs():
session.flush()
question_ids.append(str(new_q.id))
# Create Test record
test_obj = models.Test(
title=f"AI DPP: {topics[idx]}",
title=f"AI DPP: {mock['topic']}",
type="DPS",
exam_pattern=models.ExamEnum.JEE_MAIN,
sections={
@@ -98,7 +135,7 @@ async def seed_pdfs():
}
},
total_time=60,
difficulty="Hard" if idx % 2 == 0 else "Medium",
difficulty=mock["difficulty"],
scheduled_date=datetime.utcnow(),
is_active=True,
created_by=admin_id
@@ -109,4 +146,3 @@ async def seed_pdfs():
if __name__ == "__main__":
asyncio.run(seed_pdfs())

View File

@@ -113,7 +113,9 @@ export default function DPPPage() {
</Link>
) : (
<Link href={`/cbt/${dpp.id}`}>
<Button variant="secondary" size="icon" className="rounded-full group-hover:bg-[#00F0FF] group-hover:text-background transition-colors"><Play size={16} className="ml-1" /></Button>
<Button variant="secondary" size="icon" className="rounded-full group-hover:bg-[#00F0FF] group-hover:text-black dark:group-hover:text-black hover:bg-[#00F0FF] hover:text-black dark:hover:text-black shadow-sm transition-all duration-300">
<Play size={16} className="ml-1" />
</Button>
</Link>
)}
</div>