feat: improve local LLM stability with robust JSON parsing and memory…

[karpizin]

Linked Issue

Closes #

Description

• Improves local LLM stability for memory extraction, especially with Ollama-hosted models that often return noisy or non-strict JSON output.
• Makes JSON parsing more robust by extracting structured content from markdown code blocks, preamble text, and other imperfect local-model responses instead of assuming a clean raw JSON payload.
• Adds support for memory granularity via entry_mode, allowing "fact" mode for atomic memories and "chunk" mode for larger semantic segments that preserve more document context.
• Improves package import resilience by avoiding a hard failure when mem0ai package metadata is unavailable in local/dev environments.
• Updates Ollama and memory-operation docs to explain local-model optimizations and when to use chunk-based extraction.

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Refactor (no functional changes)
• Documentation update

Breaking Changes

N/A

Test Coverage

• I added/updated unit tests
• I added/updated integration tests
• I tested manually (describe below)
• No tests needed (explain why)

Manual verification:

• rebased the branch onto the latest main
• verified conflict resolution in mem0/memory/utils.py
• checked the new extract_json() behavior for:
  • fenced ```json code blocks
  • responses with leading/trailing text around JSON
  • raw JSON responses
• reviewed the new entry_mode="chunk" flow in mem0/memory/main.py and prompt/config wiring
• reviewed documentation updates for Ollama and memory add operations

Added/updated tests:

• tests/memory/test_granularity.py
• tests/memory/test_utils.py

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my code
• I have added tests that prove my fix/feature works
• New and existing tests pass locally
• I have updated documentation if needed

[CLAassistant]

All committers have signed the CLA.