fix(neuron): render HF chat templates via minijinja pycompat

The Qwen3.6 chat_template.jinja (now loaded after the precedence fix)
failed to render in minijinja: it uses Python str methods
(content.startswith/endswith/split/rstrip/lstrip) and the raise_exception
global that HF transformers patches into its Jinja env but minijinja
doesn't provide. The render error tripped the text-only fallback, so
image requests still produced zero <|image_pad|> tokens.

Wire the standard bridge into render_chat_template:
- minijinja-contrib `pycompat::unknown_method_callback` supplies the
  Python string/list/dict methods;
- a `raise_exception` global maps to a render error (so malformed inputs
  — e.g. an image in a system message — surface cleanly).

Add the real Qwen3.6-27B chat_template.jinja (verbatim from beast's HF
cache) as a test fixture and assert it renders one <|image_pad|> for a
text+image turn — the end-to-end check that would have caught this
before deploy.

Refs #16 / TP-vision.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

crates/neuron/Cargo.toml

@@ -76,15 +76,19 @@ cudarc = { version = "0.19", optional = true, default-features = false, features
 half = { version = "2.5", optional = true }
 tokenizers = { version = "0.22", default-features = false, features = ["onig"] }
 hf-hub = { version = "0.4", features = ["tokio"] }
-# Jinja-compatible template renderer for the model's
-# `tokenizer_config.json::chat_template`. Hugging Face's chat
-# templates use a strict subset of Jinja2 that minijinja supports
-# out of the box. ~80KB compiled; pure Rust, no async surface.
-# Features: `builtins` for the `is defined` / `default` filters HF
-# templates use; `json` for `tojson` (some Qwen3 templates emit
-# tool definitions via tojson); `serde` so we can hand it a
-# serde_json::Value as the context.
+# Jinja-compatible template renderer for the model's chat template
+# (standalone `chat_template.jinja` or `tokenizer_config.json::chat_template`).
+# Hugging Face's chat templates lean on Python string semantics; we
+# bridge them with `minijinja-contrib`'s `pycompat` callback (str
+# methods like `startswith`/`split`/`strip`) plus a `raise_exception`
+# global. Features: `builtins` for `is defined` / `default`; `json`
+# for `tojson`; `serde` so we can hand it a serde_json::Value context.
 minijinja = { version = "2", features = ["builtins", "json", "serde"] }
+# Python-compatibility shim: the Qwen3-VL / Qwen3.6 template uses
+# `content.startswith(...)`, `.endswith(...)`, `.split(...)`,
+# `.rstrip(...)`, `.lstrip(...)` — Python str methods minijinja doesn't
+# implement natively. `pycompat::unknown_method_callback` supplies them.
+minijinja-contrib = { version = "2", features = ["pycompat"] }
 # Direct dep on `safetensors` (re-exported by candle but its `TensorView`
 # / `slice::IndexOp` types are public-but-not-re-exported). Used by the
 # tp `fused_load` module to read per-rank slices of fused QKV tensors