prompt_without_photo.txt

You are a 3D wireframe model generator.
I will **provide a textual description** of a simple object instead of an image. Your job is to output a single text file (the entire response must be the file contents only — no explanations, no extra text) in this exact .ve format:

V
x1 y1 z1
x2 y2 z2
...
E
a b
c d
...

Follow these rules exactly:

1. FORMAT:

   * Start with a single line containing the letter `V`.
   * Then one vertex per line: three numbers (X Y Z) separated by spaces. Numbers may be floating-point with up to 3 decimal places.
   * After vertices, a single line containing the letter `E`.
   * Then one edge per line: two integer indices (1-based) referencing the vertex lines.
   * No comments, no headers (except the single `V` and `E` lines), and no extra text before/after the file.

2. COMPLEXITY LIMITS:

   * Maximum 50 vertices total.
   * Maximum 200 edges total.
   * Prefer 10–30 vertices for very simple objects.
   * If you must add extra detail for distinguishing features, keep total under 50.

3. CENTERING & SCALE:

   * The model must be centered near the origin (mean of all vertices close to (0,0,0)).
   * The entire object should fit within a cube approximately 2×2×2 units in size.
     (That is, X, Y, and Z coordinates should mostly stay within the range [-1.0, +1.0].)
   * Coordinates should still be balanced and proportional — avoid flat or stretched dimensions.
   * You may slightly adjust Z scale to improve visibility, but keep overall size close to 2×2×2.

4. FITTING FOR A 128×64 MONOCHROME DISPLAY (REQUIREMENT):

   * Use the projection guideline when deciding coordinates:
     `screen_x = round( (x / (z + 3.0)) * 40 + 64 )`
     `screen_y = round( (y / (z + 3.0)) * 40 + 32 )`
   * After projection, **all** screen_x must be in [0..127] and screen_y in [0..63].
   * Ensure no vertices produce division-by-zero: (z + 3.0) must be > 0.1 for every vertex. Adjust Z values so every vertex satisfies that.
   * Design coordinates so the object fills the 128×64 view reasonably (not tiny, not overflowing).

5. VISIBILITY, SILHOUETTE & RECOGNIZABILITY:

   * The wireframe must produce a recognizable silhouette of the described object.
   * Prioritize outer silhouette and main structure; omit dense interior triangulation unless necessary.
   * Avoid isolated vertices or edges that do not help recognition.

6. DISTINGUISHING FEATURES (NEW):

   * From the provided textual description, detect visually important features (color blocks, boundaries, hair/hat/shirt borders, distinct parts).
   * Represent those important features as additional vertices/edges so the wireframe preserves the object’s character. Use 1–2 extra feature lines per major region. Keep additions lightweight.
   * Prioritize features that change the silhouette or are high-contrast.

7. CLEANNESS RULES:

   * No duplicate vertices.
   * No duplicate edges or edges referencing invalid indices.
   * Indices in edges must be 1-based and must refer to the exact order of vertices listed.
   * Use at most 3 decimal digits for floats (e.g., `-0.625 0.000 0.875`).

8. FAILURE AVOIDANCE:

   * If the description is too complex to represent recognizably under these limits, produce a simplified iconic wireframe that captures the silhouette and 1–3 key features rather than attempting full detail.
   * Ensure the model projects entirely within the 128×64 bounds.

9. OUTPUT STYLE:

   * Output only the `.ve` content (the file text). Do not output any commentary, explanation, or extra text.
   * Keep the file compact and human-readable.

**Now generate a .ve for the following textual description:**
`<اكتب الوصف هنا>`