Academic Figure Generation
Thin CLI wrapper around PaperBanana (a.k.a. PaperVizAgent), a
multi-agent figure-generation pipeline for academic papers.
The skill provides exactly one script: scripts/generate.py. It feeds
your method text + caption into PaperBanana and writes N candidate PNGs.
Model selection and API keys come from PaperBanana's own
configs/model_config.yaml — the wrapper does not override them.
One-time setup
Clone PaperBanana somewhere convenient:
git clone https://github.com/dwzhu-pku/PaperBanana.git ~/PaperBanana
cd ~/PaperBanana
uv venv && uv pip install -r requirements.txt
Configure configs/model_config.yaml — set the image model and
the matching API key. Two common setups:
defaults:
image_model_name: "gemini-3-pro-image-preview" # or "openai/gpt-5.4-image-2"
model_name: "gemini-3.1-pro-preview" # text model for Planner/Stylist/Critic
api_keys:
google_api_key: "..." # required for Gemini models
openrouter_api_key: "" # required for openai/gpt-5.4-image-2
Use Gemini if you have a Google AI key; use GPT-Image-2 via OpenRouter
if you have an OpenRouter key. Pick one — there's nothing else to wire
up.
Workflow
Step 1: Gather inputs
You need:
- Method text: the relevant section of the paper describing the
approach (
./method.md or ./method.tex).
- Figure caption: the target caption, e.g.
"Figure 1: Overview of our framework".
If the user only gives a vague request, ask:
- What aspect of the method should the figure focus on?
- Style? (block diagram, flowchart, pipeline, architecture, comparison)
- Venue / column width? (ACL ≤ 7.5", NeurIPS single-column 5.5")
Step 2: Generate
~/PaperBanana/.venv/bin/python scripts/generate.py \
--paperbanana-root ~/PaperBanana \
--method-file ./method.md \
--caption "Figure 1: Overview of our framework" \
--out-dir ./figures/v1 \
--candidates 3 \
--aspect-ratio 16:9
| Flag |
Default |
Notes |
--paperbanana-root |
(required) |
Path to your PaperBanana checkout |
--method-file |
(required) |
Method section as a text/markdown file |
--caption |
(required) |
Target figure caption |
--out-dir |
(required) |
Where PNGs land |
--candidates |
3 |
Independent diagram candidates |
--max-concurrent |
2 |
Cap concurrent runs (be gentle on quota) |
--exp-mode |
demo_full |
Full pipeline (Planner+Stylist+Visualizer+Critic). Use demo_planner_critic to skip Stylist, or vanilla for single-shot. |
--aspect-ratio |
16:9 |
One of 21:9, 16:9, 3:2, 1:1 |
--max-critic-rounds |
2 |
Critique → revise loops (early-exits if critic says "No changes needed") |
Step 3: Present & iterate
- Show all candidates to the user.
- Common refinements: color scheme, layout, label text, font size.
- Re-run with a tweaked caption or more candidates.
Step 4: Export
- PNGs are written as
candidate_0.png, candidate_1.png, … in --out-dir.
- For camera-ready PDFs:
magick candidate_0.png candidate_0.pdf.
Style guidelines
- Color: consistent, colorblind-friendly palette
- Fonts: match the paper's body font (Times for ACL/EMNLP,
Helvetica/Arial for many ML venues)
- Labels: concise; no full sentences inside the diagram
- Arrows: solid for data flow, dashed for optional / feedback loops
- Whitespace: don't overcrowd — reviewers skim figures in seconds
Common figure types
| Type |
When to use |
Key elements |
| Pipeline / Flowchart |
Sequential processing |
Boxes + arrows, L→R or T→B |
| Architecture |
System overview |
Nested boxes, clear module boundaries |
| Comparison |
Before/after, baseline vs proposed |
Side-by-side panels |
| Ablation |
Component contributions |
Bar charts, highlighted rows |
| Framework |
High-level conceptual overview |
Abstract shapes, minimal detail |
Troubleshooting
429 RESOURCE_EXHAUSTED on Gemini: monthly Google AI Studio
spending cap hit. Raise it at https://ai.studio/spend or switch
image_model_name to openai/gpt-5.4-image-2 and set
OPENROUTER_API_KEY.OpenRouter Client not initialized: OPENROUTER_API_KEY not in env
and openrouter_api_key not in yaml.- No PNGs in output dir: check
out_dir/results.json for the raw
per-candidate response and any error messages.
- Long latency (>5 min): most wall time is the image model. Lower
--candidates or use --exp-mode vanilla for faster iteration.
Links
1---2name: academic-figure-generation3description: Generates publication-quality academic figures (framework diagrams, pipeline illustrations, system architectures, method overviews) from a paper's method text and a target caption, using a local PaperBanana multi-agent pipeline (Retriever → Planner → Stylist → Visualizer → Critic).4---56# Academic Figure Generation78Thin CLI wrapper around **PaperBanana** (a.k.a. PaperVizAgent), a9multi-agent figure-generation pipeline for academic papers.1011The skill provides exactly one script: `scripts/generate.py`. It feeds12your method text + caption into PaperBanana and writes N candidate PNGs.13Model selection and API keys come from PaperBanana's own14`configs/model_config.yaml` — the wrapper does not override them.1516## One-time setup17181. **Clone PaperBanana** somewhere convenient:1920 ```bash21 git clone https://github.com/dwzhu-pku/PaperBanana.git ~/PaperBanana22 cd ~/PaperBanana23 uv venv && uv pip install -r requirements.txt24 ```25262. **Configure `configs/model_config.yaml`** — set the image model and27 the matching API key. Two common setups:2829 ```yaml30 defaults:31 image_model_name: "gemini-3-pro-image-preview" # or "openai/gpt-5.4-image-2"32 model_name: "gemini-3.1-pro-preview" # text model for Planner/Stylist/Critic3334 api_keys:35 google_api_key: "..." # required for Gemini models36 openrouter_api_key: "" # required for openai/gpt-5.4-image-237 ```3839 Use Gemini if you have a Google AI key; use GPT-Image-2 via OpenRouter40 if you have an OpenRouter key. Pick one — there's nothing else to wire41 up.4243## Workflow4445### Step 1: Gather inputs4647You need:48491. **Method text**: the relevant section of the paper describing the50 approach (`./method.md` or `./method.tex`).512. **Figure caption**: the target caption, e.g. `"Figure 1: Overview of52 our framework"`.5354If the user only gives a vague request, ask:5556- What aspect of the method should the figure focus on?57- Style? (block diagram, flowchart, pipeline, architecture, comparison)58- Venue / column width? (ACL ≤ 7.5", NeurIPS single-column 5.5")5960### Step 2: Generate6162```bash63~/PaperBanana/.venv/bin/python scripts/generate.py \64 --paperbanana-root ~/PaperBanana \65 --method-file ./method.md \66 --caption "Figure 1: Overview of our framework" \67 --out-dir ./figures/v1 \68 --candidates 3 \69 --aspect-ratio 16:970```7172| Flag | Default | Notes |73|------|---------|-------|74| `--paperbanana-root` | (required) | Path to your PaperBanana checkout |75| `--method-file` | (required) | Method section as a text/markdown file |76| `--caption` | (required) | Target figure caption |77| `--out-dir` | (required) | Where PNGs land |78| `--candidates` | `3` | Independent diagram candidates |79| `--max-concurrent` | `2` | Cap concurrent runs (be gentle on quota) |80| `--exp-mode` | `demo_full` | Full pipeline (Planner+Stylist+Visualizer+Critic). Use `demo_planner_critic` to skip Stylist, or `vanilla` for single-shot. |81| `--aspect-ratio` | `16:9` | One of `21:9`, `16:9`, `3:2`, `1:1` |82| `--max-critic-rounds` | `2` | Critique → revise loops (early-exits if critic says "No changes needed") |8384### Step 3: Present & iterate8586- Show all candidates to the user.87- Common refinements: color scheme, layout, label text, font size.88- Re-run with a tweaked caption or more candidates.8990### Step 4: Export9192- PNGs are written as `candidate_0.png`, `candidate_1.png`, … in `--out-dir`.93- For camera-ready PDFs: `magick candidate_0.png candidate_0.pdf`.9495## Style guidelines9697- **Color**: consistent, colorblind-friendly palette98- **Fonts**: match the paper's body font (Times for ACL/EMNLP,99 Helvetica/Arial for many ML venues)100- **Labels**: concise; no full sentences inside the diagram101- **Arrows**: solid for data flow, dashed for optional / feedback loops102- **Whitespace**: don't overcrowd — reviewers skim figures in seconds103104## Common figure types105106| Type | When to use | Key elements |107|------|-------------|--------------|108| Pipeline / Flowchart | Sequential processing | Boxes + arrows, L→R or T→B |109| Architecture | System overview | Nested boxes, clear module boundaries |110| Comparison | Before/after, baseline vs proposed | Side-by-side panels |111| Ablation | Component contributions | Bar charts, highlighted rows |112| Framework | High-level conceptual overview | Abstract shapes, minimal detail |113114## Troubleshooting115116- **`429 RESOURCE_EXHAUSTED` on Gemini**: monthly Google AI Studio117 spending cap hit. Raise it at <https://ai.studio/spend> or switch118 `image_model_name` to `openai/gpt-5.4-image-2` and set119 `OPENROUTER_API_KEY`.120- **`OpenRouter Client not initialized`**: `OPENROUTER_API_KEY` not in env121 and `openrouter_api_key` not in yaml.122- **No PNGs in output dir**: check `out_dir/results.json` for the raw123 per-candidate response and any error messages.124- **Long latency (>5 min)**: most wall time is the image model. Lower125 `--candidates` or use `--exp-mode vanilla` for faster iteration.126127## Links128129- PaperBanana repo: <https://github.com/dwzhu-pku/PaperBanana>130- PaperVizAgent (Google Research version of the same project): <https://github.com/google-research/papervizagent>
jxtse