Uploaded complete moirai_forecast_app contents

README.md

@@ -121,6 +121,14 @@ worth recording because the diagnosis process matters as much as the fix:
    prebuilt wheel instead of a source build. No requirements.txt change
    was actually needed — my earlier "fix" there was chasing the wrong
    variable, corrected here for the record rather than quietly erased.
+4. **Third failure, at runtime this time**: the build succeeded (scipy
+   installed as a wheel), but `import gradio` crashed with
+   `ImportError: cannot import name 'HfFolder' from 'huggingface_hub'`.
+   `huggingface_hub` 1.0 removed the long-deprecated `HfFolder` class;
+   `gradio==4.44.0`'s `oauth.py` still imports it unconditionally.
+   `uni2ts` only floors `huggingface_hub>=0.23.0` with no upper bound,
+   so pip picked the newest 1.x. **Fix**: `requirements.txt` now pins
+   `huggingface_hub>=0.23.0,<1.0`.
 
 ## Architecture
 

requirements.txt

@@ -3,6 +3,15 @@ uni2ts>=2.0.0
 torch>=2.1
 gluonts>=0.14
 
+# Pinned explicitly: huggingface_hub 1.0+ removed the HfFolder class,
+# but gradio 4.44.x's oauth.py still does `from huggingface_hub import
+# HfFolder, whoami` unconditionally at import time -- so `import gradio`
+# crashes immediately on any resolution that picks huggingface_hub>=1.0.
+# uni2ts only floors huggingface_hub>=0.23.0 with no ceiling, so without
+# this pin pip is free to grab the latest 1.x. Confirmed against a real
+# Space's runtime traceback, not a guess.
+huggingface_hub>=0.23.0,<1.0
+
 # Data
 yfinance>=0.2.40
 pandas>=2.0