Python: Fix broken samples and add missing READMEs

[giles17]

Motivation and Context

Fix runtime errors in several Python samples and add missing README documentation for sample folders.

Description

Sample fixes

• simple_context_provider.py: instructions was passed as a top-level kwarg to get_response(), but it must be inside the options dict. This caused a silent failure (swallowed by suppress(Exception)) where user info extraction never worked.
• suspend_resume_session.py: The in-memory session demo used FoundryChatClient() with no args, which requires a credential. Changed to OpenAIChatCompletionClient() to match the docstring intent of comparing service-managed vs in-memory sessions.
• foundry_chat_client_with_hosted_mcp.py: store=True was passed as a top-level kwarg to Agent.run(), but it must be inside options={"store": True}. This caused a TypeError crash.
• mem0_basic.py: Preferences query was never sent to the agent before the Mem0 indexing delay, so Mem0 never learned user preferences. Also added missing print for new session result.
• mem0_sessions.py: Agent-scoped example had no session for multi-turn conversation and combined user_id with agent_id which Mem0 API does not support (memories stored without user_id when agent_id is provided).
• redis_basics.py: Standalone provider demo printed raw SessionContext object instead of retrieved context messages. Fixed docstring referencing wrong client type.
• redis_sessions.py: Missing load_dotenv() call caused KeyError on FOUNDRY_PROJECT_ENDPOINT. Fixed docstring referencing wrong client type.
• azure_redis_conversation.py: Duplicate copyright comment replaced with missing load_dotenv() call.

New READMEs

• context_providers/README.md: Top-level README covering the two standalone samples and linking to subfolders (azure_ai_search, mem0, redis) that already have their own READMEs.
• conversations/README.md: README covering suspend/resume sessions, custom history provider, and Redis history provider samples with prerequisites.

Checklist

• I've read the contributing guidelines
• I've verified changes locally

[Copilot]

Pull request overview

Fixes several Python sample runtime issues caused by incorrect keyword argument placement, and adds missing README documentation for sample folders.

Changes:

• Fix Agent.run() and get_response() option passing by nesting previously top-level kwargs into options={...}.
• Update the in-memory session example to use OpenAIChatCompletionClient() instead of FoundryChatClient() to match the sample’s intent.
• Add top-level READMEs for conversations/ and context_providers/ sample folders.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
python/samples/02-agents/providers/foundry/foundry_chat_client_with_hosted_mcp.py	Fixes incorrect store=True kwarg usage by moving it into options for Agent.run().
python/samples/02-agents/conversations/suspend_resume_session.py	Switches the in-memory demo client to OpenAIChatCompletionClient() to align with the sample’s described comparison.
python/samples/02-agents/conversations/README.md	Adds documentation for conversation/session samples and per-sample prerequisites.
python/samples/02-agents/context_providers/simple_context_provider.py	Fixes incorrect instructions= placement by moving it into the options dict for get_response().
python/samples/02-agents/context_providers/README.md	Adds documentation for context provider samples and links to subfolder READMEs.