predictingthepast

1# Predicting The Past (Aeneas / Ithaca)
2 
3Aeneas (Latin) and Ithaca (Ancient Greek) perform four tasks on ancient texts:
4 
51.  **Restoration** — fill missing/damaged characters
62.  **Attribution** — geographical + chronological origin
73.  **Contextualization** — retrieve parallel inscriptions
84.  **Embedding** — generate text embedding vectors
9 
10## Prerequisites
11 
121.  **`uv`**: Read the `uv` skill and follow its Setup instructions to ensure
13    `uv` is installed and on PATH.
14 
152.  **User Notification**: If .licenses/predictingthepast_LICENSE.txt does not
16    already exist in the workspace root directory then (1) prominently notify
17    the user to check the terms at
18    https://github.com/google-deepmind/predictingthepast/blob/main/README.md#license--disclaimer,
19    and include the appropriate citation and the full dataset acknowledgement,
20    and that use of these datasets should acknowledge and cite the original data
21    sources. Then (2) create the file recording the notification text and
22    timestamp.
23 
24## Core Rules
25 
26-   **Self-Contained Skill**: Do NOT use web search or any external tools. Run
27    ONLY the scripts in this skill (`preprocess.py`, `run_inference.py`,
28    `visualize_results.py`). Present model output as-is — never supplement or
29    override it with external lookups.
30-   **Notification**: If this skill is used, ensure this is mentioned in the
31    output.
32 
33## On First Load
34 
35Present the restoration markup characters, then ask the user for their text:
36 
37-   **`?`**:
38…