[Code scan] Airflow example contains stale local paths and DAG assumptions

Source: Codex global repository scan of deepmodeling/dpti at commit b719828e7eeb571bd26411430197cd74ea78e38c.
Project: https://github.com/orgs/deepmodeling/projects/27

Problem
The Airflow example hard-codes a local user path and is not aligned with the README wording. It points to `TI_taskflow` and a `work_base_dir` under `/home/fengbo/...`, so copying it on a clean machine will fail unless the user rewrites private paths and has the matching DAG loaded.

Code references
https://github.com/deepmodeling/dpti/blob/b719828e7eeb571bd26411430197cd74ea78e38c/examples/airflow.sh#L1
https://github.com/deepmodeling/dpti/blob/b719828e7eeb571bd26411430197cd74ea78e38c/examples/FreeEnergy.json#L6
https://github.com/deepmodeling/dpti/blob/b719828e7eeb571bd26411430197cd74ea78e38c/README.md#L177

Reproduction
Try to run `examples/airflow.sh` in a clean checkout with a normal Airflow setup.

Expected result
Examples should avoid private absolute paths and should name the DAG that the package actually installs or documents for that workflow.

Inspect examples/airflow.sh, examples/FreeEnergy.json, and the README section linked in the issue. Reproduce the workflow in a clean Airflow checkout, then verify that the example avoids private absolute paths and names the DAG used by the documented workflow.

Written by the indexing model from the issue text.