chore: remove DashScope placeholder defaults, sync docs, and clean up debug prints

[neverbiasu]

Summary

• Replaced hardcoded DashScope placeholder API keys with None defaults across the codebase.
• Updated the demo script to read the key from the DASHSCOPE_API_KEY environment variable.
• Synced documentation fixes: corrected CLI flags, requirements naming, and added missing download commands.
• Removed leftover debug print blocks from the scheduler.

Changes

DashScope API key defaults

• inference.py, inference_turbo.py: changed --dashscope_api_key argparse default from "sk-xxxxxxxxxxxxxxxxxxxxxxxxxx" to None, with an updated help string.
• boogu/pipelines/boogu/pipeline_boogu.py: changed 4 dashscope_api_key parameter defaults from the placeholder string to None (type updated to Optional[str]). The existing assertion that rejects the placeholder is preserved.
• demo_scripts/demo_ti2i_remote_reasoning.sh: replaced the hardcoded placeholder key with a check that reads DASHSCOPE_API_KEY from the environment, exiting with a helpful error message if it is not set.

Documentation sync

• README.md, README_CN.md: fixed --model → --pretrained_pipeline_name_or_path, corrected requirements file naming convention (<torch>-<cuda>.txt), added missing Edit-Turbo download command (CN), and documented the auto-generated conda environment name suffix.
• INFERENCE_GUIDE.md: added Turbo entry-point mapping (inference.py for Base/Edit, inference_turbo.py for Turbo/Edit-Turbo) and a minimal Turbo CLI example.

Debug cleanup

• boogu/schedulers/scheduling_flow_match_euler_discrete_time_shifting.py: removed two commented-out ######debug###### blocks containing leftover print statements.

[neverbiasu]

@Boogu-Team Hi, could you review this please.