[Example] ggml: support experimental embedding

[hydai]

Related: https://github.com/WasmEdge/WasmEdge/pull/3189

[juntao]

Hello, I am a code review bot on flows.network. Here are my reviews of code commits in this PR.

---

Commit 35d8aec1b333315bb1f513a3ea43f2bee2286438

Summary of key changes:

• Added new files: Cargo.toml, README.md, and src/main.rs.
• Added a new binary called wasmedge-ggml-llama-embedding.
• Added dependencies to Cargo.toml, including serde_json and wasi-nn with a specified branch.
• Added a new directory called wasmedge-ggml-llama-embedding which contains the new files.

Potential problems:

• It's unclear from the patch what the purpose of this code is, and what problem it solves.
• There are no unit tests included in the patch, which is a concern for the reliability and correctness of the code.
• The patch includes binary and data files (wasmedge-ggml-llama-embedding.wasm and llama-2-7b-chat.Q5_K_M.gguf) without any explanation of how they are used or where they come from. This could be a violation of licensing or intellectual property rules.
• The code includes a mix of console input/output and environment variables, which could make it difficult to understand and use.
• The code is missing error handling and proper exception handling, which can lead to unexpected behavior or crashes.
• The code uses hard-coded paths ($HOME/.zshenv and $HOME/.bashrc) which may not be portable across different environments.
• The code does not follow consistent code formatting (e.g., inconsistent indentation, unnecessary newlines).

Overall, the patch lacks important context, documentation, and testing, and may have potential issues with portability, licensing, and error handling. It would be beneficial to address these concerns before merging the pull request.