grep -n -i "thorax\|chest wall\|rib cage\|sternum\|intercostal\|pleura\|mediasti…

"Thorax" in /textbooks/9780323547048

You are building a detailed thorax anatomy PDF from Netter's Atlas of Human Anatomy (ISBN: 9780323547048). The textbook block file is at `/textbooks/9780323547048/9780323547048_block1.md`. The thorax section starts around line/offset 1800-2000 of this file (pages ~300-450 in the book). Your tasks: 1. Read the file starting at offset 1700, reading 400 lines at a time, to find ALL thorax-related sections. Key sections to find include: - Thoracic wall / chest wall / ribs / sternum / intercostal spaces - Breast / mammary gland - Pleura and lungs - Heart and pericardium - Mediastinum - Diaphragm - Thoracic vessels (aorta, vena cava, pulmonary vessels) - Thoracic nerves (vagus, phrenic, sympathetic trunk) - Esophagus / trachea 2. For EACH section found, collect: - The heading text - The page number (from <!-- page:XXX --> markers) - The image URLs (all the cdn.orris.care image links, including their alt text descriptions) - Any explanatory text paragraphs present 3. Read multiple chunks (offset 1700, 2100, 2500, 2900, 3300, 3700, 4100, 4500) to cover the full thorax coverage. Stop when you reach abdominal content. 4. Return a structured JSON list of sections, each with: - section_name - page_number - images: list of {url, description} - text: any prose text Return ALL image URLs found in the thorax sections - these are the Netter diagrams that will appear in the PDF.

"## " in /textbooks/9780323547048

You are extracting thorax anatomy content from two medical textbooks in a library database. Your goal is to return a structured JSON with all thorax sections, their text, and image URLs. ## BOOK 1: Netter's Atlas of Human Anatomy (ISBN: 9780323547048) File: `/textbooks/9780323547048/9780323547048_block1.md` The thorax section in Netter starts around page 155-250 (file offset ~1800-2600). Read the file in chunks of 300 lines at these offsets: 1700, 2000, 2300, 2600, 2900. For each chunk, collect: - Every `## Heading` text found - Every `<!-- page:XXX -->` page number - Every image line like `` — capture BOTH the alt text description AND the URL - Any prose/text paragraphs (non-image, non-heading lines with actual content) STOP collecting when you see abdominal content (stomach, colon, liver, kidney, etc.) appearing repeatedly. ## BOOK 2: Color Atlas of Human Anatomy Vol 1 (ISBN: 9783132424432) File: `/textbooks/9783132424432/9783132424432_block1.md` The "2 Trunk" section starts at line 142. Read from offset 142 onwards in chunks of 300 lines (offsets: 142, 442, 742, 1042) to extract thorax content including: - Thoracic wall, ribs, sternum, intercostal spaces - Pleura and lungs - Heart and pericardium - Mediastinum, vessels, nerves For each image found, capture the URL and alt text. For each section, capture all prose text. ## OUTPUT FORMAT Return a JSON object with two keys: 1. "netter_thorax": list of { "page": number, "heading": string, "images": [{"url": string, "description": string}], "text": string } 2. "color_atlas_thorax": list of { "heading": string, "images": [{"url": string, "description": string}], "text": string } Include ALL images found — these are critical for the final PDF. Return as much detail as possible.