Initial commit with LFS-tracked PDFs

.gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+*.pdf filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,3 @@
+.env
+venv
+__pycache__/

Dockerfile

@@ -0,0 +1,16 @@
+# Use official Python image
+FROM python:3.10-slim
+
+RUN apt-get update && apt-get install -y git
+
+RUN useradd -m -u 1000 user
+USER user
+ENV PATH="/home/user/.local/bin:$PATH"
+
+WORKDIR /app
+
+COPY --chown=user ./requirements.txt requirements.txt
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
+
+COPY --chown=user . /app
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

LearningPointsExtractor.md

@@ -0,0 +1,77 @@
+# PDF Learning Points Extractor
+
+You are an expert at extracting key learning points from educational PDFs and presenting them as bite-sized, actionable insights perfect for mobile notifications and quick learning.
+
+## Your Task
+
+When given a PDF document, extract the most important, practical, and memorable points that would help someone learn the subject matter effectively. Each point should be:
+
+1. **Self-contained**: Understandable without additional context
+2. **Actionable**: Something the learner can immediately understand or apply
+3. **Concise**: Brief enough to read in a notification (2-3 sentences max)
+4. **Valuable**: Represents a key concept, principle, or insight from the material
+
+## Output Format
+
+Return ONLY a valid JSON object in this exact format:
+
+```json
+{
+  "title": "[Clear, descriptive title - max 8 words]",
+  "content": "[The important point/statement - 2-3 sentences max, clear and concise]"
+}
+```
+
+## Guidelines
+
+- **Title**: Should be specific and descriptive (e.g., "Stack Memory Management" not just "Memory")
+- **Content**: Should explain the concept clearly, include practical relevance when possible
+- Extract ONE point per request - make it the most impactful point you can find
+- Prioritize foundational concepts, practical techniques, and key insights
+- Avoid filler words - be direct and clear
+- Use simple language while maintaining technical accuracy
+- If the PDF covers multiple topics, focus on core principles first
+
+## Important Notes
+
+- Extract points sequentially through the document for comprehensive coverage
+- Focus on concepts that have lasting value, not just facts
+- Ensure each point teaches something meaningful
+- Keep the learner engaged with clear, practical insights
+- Always return valid, parseable JSON
+
+## Ensuring Uniqueness with Date-Based Page Selection
+
+**CRITICAL**: You will extract a DIFFERENT point each day using date-based page rotation.
+
+### Page Selection Algorithm
+
+1. The user will provide:
+   - Current date (e.g., "2025-10-02")
+   - Total number of pages in the PDF (e.g., 250)
+
+2. Calculate the target page using modulo:
+   ```
+   Day of Year = Calculate from the date (1-365/366)
+   Target Page = (Day of Year % Total Pages) + 1
+   ```
+   If Day of Year is 275 and Total Pages is 250:
+   Target Page = (275 % 250) + 1 = 26
+
+3. Extract the most important/interesting learning point from that specific page
+
+4. If the calculated page has no substantial content (cover page, blank, TOC):
+   - Move to the next page with actual content
+   - Mention in your response which page you used
+
+### Your Process
+
+1. Calculate: Day of year = 275
+2. Calculate: Target page = (275 % 250) + 1 = 26
+3. Extract the best learning point from page 26
+4. If page 26 is blank/non-content, use the nearest content page
+5. Return the JSON with the point from that page
+
+**Note**: Clearly identify which page number you extracted from in your internal process.
+
+When ready, calculate the target page and extract one significant learning point from that specific page.

ProgrammingGroundUp-1-0-booksize.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:663bd554622af154a94e0363fbd8b5b3e93137247f6eeada77005c911ec74513
+size 1383853

README.md

@@ -0,0 +1,11 @@
+---
+title: CruxNow
+emoji: ⚡
+colorFrom: purple
+colorTo: purple
+sdk: docker
+pinned: false
+short_description: Smart updates. Core insights. Delivered instantly.
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

The C Programming Language (Kernighan Ritchie).pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94ed541af52448918b11ac6fb257a6589903ab6340c61c2713d1e1b7be0e3a68
+size 1143598

app.py

@@ -0,0 +1,83 @@
+from fastapi import FastAPI
+from gemiwrap import GeminiWrapper
+from google import genai
+from google.genai import types
+import json_repair
+from functools import partial
+import os
+if os.path.exists(".env"):
+	from dotenv import load_dotenv
+	load_dotenv()
+
+app = FastAPI()
+
+geminiWrapper = partial(GeminiWrapper,
+	model_name="gemini-flash-lite-latest",
+	schema=genai.types.Schema(
+		type = genai.types.Type.OBJECT,
+		required = ["title", "content"],
+		properties = {
+			"title": genai.types.Schema(
+				type = genai.types.Type.STRING,
+			),
+			"content": genai.types.Schema(
+				type = genai.types.Type.STRING,
+			),
+		},
+	),
+	delete_files=True
+)
+
+@app.get("/")
+def greet_json():
+	return {"Hello": "World!"}
+
+@app.get("/breaking_news")
+def breaking_news():
+	user_prompt = None
+	with open("breaking_news.md", 'r') as file:
+		user_prompt = file.read()
+
+	grounding_tool = types.Tool(
+		google_search=types.GoogleSearch()
+	)
+	model_responses = geminiWrapper(tools=[grounding_tool], response_mime_type="text/plain").send_message(user_prompt=user_prompt)
+	return json_repair.loads(model_responses[0])
+
+@app.get("/hacker_news")
+def hacker_news():
+	with open("test.txt", 'w') as file:
+		file.write("Hey, Hey")
+
+	text = None
+	with open("test.txt", 'r') as file:
+		text = file.read()
+	return {"Hello": text}
+
+@app.get("/hackers_law")
+def hackers_law():
+	user_prompt = None
+	with open("hackers_law.md", 'r') as file:
+		user_prompt = file.read()
+
+	model_responses = geminiWrapper().send_message(user_prompt=user_prompt)
+	return json_repair.loads(model_responses[0])
+
+@app.get("/pdf_crux")
+def pdf_crux(name: str):
+	system_prompt = None
+	with open("LearningPointsExtractor.md", 'r') as file:
+		system_prompt = file.read()
+	# name : VonNeumann, hacker_laws,
+	file_path = "ProgrammingGroundUp-1-0-booksize.pdf"
+	if name == "assembly":
+		file_path = "ProgrammingGroundUp-1-0-booksize.pdf"
+	elif name == "arm":
+		file_path = "arm-baremetal-ebook.pdf"
+	elif name == "cpumemory":
+		file_path = "cpumemory.pdf"
+	elif name == "c":
+		file_path = "The C Programming Language (Kernighan Ritchie).pdf"
+
+	model_responses = geminiWrapper().send_message(user_prompt="", system_instruction=system_prompt, file_path=file_path)
+	return json_repair.loads(model_responses[0])

arm-baremetal-ebook.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2da89ad0d2e1fcef470c046ad57d5904b9076c2741606862b8b7710294c461c
+size 692678

breaking_news.md

@@ -0,0 +1,48 @@
+# Breaking News Finder - Enhanced Prompt
+
+## Task
+Search for and report the most significant breaking news story from the past 24 hours in one of these categories: World, Politics, Business, Technology, Science, Health, Entertainment, or Sports.
+
+## Selection Criteria
+- **Recency**: Published within the last 24 hours
+- **Significance**: Major developments that would lead news broadcasts or front pages
+- **Verification**: From established, credible news organizations only
+- **Impact**: Stories affecting large populations, markets, or having widespread consequences
+
+## Source Requirements
+Prioritize news from these types of sources:
+- Major international news agencies (Reuters, AP, BBC, CNN, etc.)
+- Established newspapers and news websites
+- Government or official organizational announcements
+- Verified social media accounts of news organizations
+
+## Content Guidelines
+- Use objective, professional journalism language
+- Include specific factual details: WHO, WHAT, WHEN, WHERE, WHY
+- Mention exact times, locations, and key figures when available
+- Focus on confirmed facts, not speculation or analysis
+- Keep content concise but comprehensive (100-200 words)
+
+## Output Format
+Return results in this exact JSON structure:
+```json
+{
+  "category": "category_name",
+  "title": "Compelling, specific headline that captures the breaking news",
+  "content": "Professional news summary with key facts, figures, and context. Include specific details about what happened, who is involved, when it occurred, and why it matters."
+}
+```
+
+## Quality Checklist
+Before finalizing, ensure the story:
+- [ ] Is genuinely breaking news (not just recently published old news)
+- [ ] Comes from a reputable, verifiable source
+- [ ] Contains specific, factual details
+- [ ] Would be considered significant by major news outlets
+- [ ] Is written in clear, professional language
+
+## Important Notes
+- Return ONLY the JSON format with no additional text
+- Do not create, embellish, or speculate on any information
+- If no genuinely breaking news is found, search for the most significant recent story that meets the criteria
+- Verify information accuracy before including in the response

cpumemory.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2902dddcadb1ec97eeabd338bfaabf80f1fa2ff4e4d769b4052da88bfbb20387
+size 934051

hackers_law.md

@@ -0,0 +1,37 @@
+# Hacker Laws Fetcher Prompt
+
+Extract one random law from the Hacker Laws website and format it for iOS notification display.
+
+## Instructions:
+1. Fetch content from: https://hacker-laws.com
+2. Select ONE law randomly from the available laws on the page
+3. Extract the law's title and its main content/description
+4. Format the output as valid JSON with "title" and "content" fields
+5. Keep the content concise but complete - suitable for iOS notification display
+6. Remove any markdown formatting, links, or extra whitespace
+7. Ensure the content is readable and fits well in a notification
+
+## Output Format:
+```json
+{
+  "title": "[Law Title]",
+  "content": "[Law Description/Content]"
+}
+```
+
+## Example Output:
+```json
+{
+  "title": "Murphy's Law",
+  "content": "Anything that can go wrong will go wrong."
+}
+```
+
+## Additional Requirements:
+- Use only plain text (no markdown, HTML, or special formatting)
+- Keep total length under 200 characters if possible for notification readability
+- If the law description is too long, summarize the core concept
+- Ensure the four dashes (----) separator is exactly as shown
+- Return valid JSON format only - no additional text or explanation
+- Use proper JSON syntax with quoted keys and values
+- Choose a different law each time if possible

requirements.txt

@@ -0,0 +1,5 @@
+fastapi
+uvicorn[standard]
+git+https://github.com/jebin2/gemiwrap.git
+google-generativeai
+json_repair