Support with_parameter and with_parameters in query engine

[leiyuou]

Support $param placeholder in query engine.
Supported Scalar Types:

Null
Boolean
Integer
Float
String

Supported Complex Types:

Float Vectors: Lists of numbers (e.g., [0.1, 0.2, 0.3]) are supported and converted to VectorLiteral (as Vec).

Key Changes:

1. Stored user parameters in a HashMap<String, serde_json::Value> within the
   CypherQuery
2. Propagated parameters to the semantic analysis layer, where they are resolved and validated.
3. Implemented a parameter substitution pass (parameter_substitution.rs) that replaces parameter placeholders (e.g., $name) with their corresponding user-provided literal values (Scalar or Vector) directly in the AST before logical planning.
4. Added end-to-end tests in tests/test_parameter_examples.rs (in plan) and verified existing tests to ensure correct parsing and execution of parameterized queries.

Closes #103.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 83.43373% with 55 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
crates/lance-graph/src/parameter_substitution.rs	79.22%	43 Missing ⚠️
crates/lance-graph/src/parser.rs	84.84%	5 Missing ⚠️
...s/lance-graph/src/datafusion_planner/expression.rs	0.00%	3 Missing ⚠️
crates/lance-graph/src/query.rs	86.66%	2 Missing ⚠️
crates/lance-graph/src/semantic.rs	97.29%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!

[ChunxuTang]

@leiyuou Thanks for the contribution! This is a very nice work.

I carefully re-evaluated the three design options for the parameter support, and I think it's better to go with option 1 (replacing the parameters during the semantic analysis pass). Could you update the PR?

[ChunxuTang]

Looks good! Thanks for the update.
Just left some minor comments (mostly for code consistency).

[ChunxuTang]

@leiyuou Generally LGTM!

During the review, I found that the python_to_json method on the Python side actually doesn't support Python lists, so such a query would fail on the Python side:

query = (
    CypherQuery("MATCH (n:Item) WHERE vector_distance(n.embedding, $q, 'cosine') < 0.5 RETURN n")
    .with_config(config)
    .with_parameter("q", [0.1, 0.2, 0.3])  # ValueError raised here
)

Could you send a follow-up PR to fix the issue?