docs(langchain): fix SummarizationMiddleware AND vs OR trigger logic documentation

[Prince (PrinceThummar011)]

Summary

Fixes #2027

The SummarizationMiddleware documentation under Built-in middleware →
Configuration options contained misleading and incorrect information about
the trigger parameter's AND vs OR logic, causing user confusion.

Problem

The Python "Full example" accordion labelled a list of tuples as
"Single condition (AND logic)", then showed a second example — also a
list of tuples — labelled "Multiple conditions (OR logic)". Both
examples were syntactically identical, making it impossible to understand
how AND logic differed from OR logic.

Additionally, the JavaScript trigger ParamField described AND/OR behaviour
in abstract text only, with no code examples to illustrate the difference.

Changes

src/oss/langchain/middleware/built-in.mdx

• Python trigger ParamField — added a <Note> block explicitly
  stating that Python's ContextSize is a single ("key", value) tuple and
  does not support AND logic. A list of tuples always evaluates as OR
  logic (summarization fires when any one condition is met).
• JavaScript trigger ParamField — added inline TypeScript code examples
  for both AND logic (trigger: { tokens: 4000, messages: 10 }) and OR logic
  (trigger: [{ tokens: 3000 }, { messages: 6 }]), making the structural
  difference immediately visible.
• "Full example" accordion (Python bullet list) — added a "No AND
  logic" bullet clarifying there is no Python equivalent of JavaScript's
  multi-property AND object.

Type of change

• Documentation fix (incorrect / misleading content)

Checklist

• Targets the correct file (src/oss/langchain/middleware/built-in.mdx)
• No changes to reference/ (auto-generated)
• docs.json navigation unchanged (no new pages added)
• Frontmatter intact on all edited files
• Follows Google Developer Documentation Style Guide