From e0c1fd9a417a8fb52ddfa715bbaddff8a1be7a8b Mon Sep 17 00:00:00 2001
From: lukasWuttke <54042461+LukasWodka@users.noreply.github.com>
Date: Sat, 29 Aug 2026 13:49:06 +0200
Subject: [PATCH] docs(sdk): remove the tensorflow extra, which does not exist
 (backend#2671) (#154)

Production docs told users to run pip install "tracebloc[tensorflow]". The SDK
publishes no such extra -- 1.0.2 has pytorch, sklearn, xgboost, catboost,
lightgbm, lifelines, scikit-survival, all and dev.

This is not cosmetic. pip WARNS and exits 0 on an unknown extra, so the command
appears to succeed while installing the core SDK only, and the failure surfaces
much later as a missing framework.

Second defect, found while fixing the first: `all` was annotated "all
frameworks" and "everything". It resolves to pytorch, sklearn, lifelines and
scikit-survival -- it does NOT include TensorFlow. Both annotations corrected to
name what it actually pulls in.

The SDK does handle TensorFlow in code (weight_loaders, rewriter, loss,
augmentation, safetensors_io), it simply is not packaged as an extra. The Note
states that fact and stops there rather than inventing an install recipe that
nobody has verified end to end.

check-sdk-extras.py: 2 problems -> 0, exit 0.

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
---
 join-use-case/start-training.mdx |  3 +--
 tools-help/tracebloc.mdx         | 15 +++++++++++++--
 2 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/join-use-case/start-training.mdx b/join-use-case/start-training.mdx
index 5f8f3b4..65b3b3d 100644
--- a/join-use-case/start-training.mdx
+++ b/join-use-case/start-training.mdx
@@ -43,9 +43,8 @@ python -m pip install --upgrade pip
 
 # Install with the extra that matches your framework:
 pip install "tracebloc[pytorch]>=0.18.1"
-# pip install "tracebloc[tensorflow]>=0.18.1"
 # pip install "tracebloc[sklearn]>=0.18.1"
-# pip install "tracebloc[all]>=0.18.1"  # all frameworks
+# pip install "tracebloc[all]>=0.18.1"  # pytorch + sklearn + lifelines + scikit-survival
 ```
 
 ## Install and Launch Jupyter Notebook
diff --git a/tools-help/tracebloc.mdx b/tools-help/tracebloc.mdx
index 2729f4a..f1d8642 100644
--- a/tools-help/tracebloc.mdx
+++ b/tools-help/tracebloc.mdx
@@ -17,13 +17,24 @@ Pick the extra that matches your ML framework — the default install ships the
 
 ```bash
 pip install "tracebloc[pytorch]>=0.18.1"                       # most users
-# pip install "tracebloc[tensorflow]>=0.18.1"                  # TensorFlow
 # pip install "tracebloc[sklearn]>=0.18.1"                     # scikit-learn, incl. XGBoost / CatBoost / LightGBM
 # pip install "tracebloc[xgboost]>=0.18.1"                     # one boosting backend only (or: catboost, lightgbm)
 # pip install "tracebloc[lifelines,scikit-survival]>=0.18.1"   # survival analysis
-# pip install "tracebloc[all]>=0.18.1"                         # everything
+# pip install "tracebloc[all]>=0.18.1"                         # pytorch + sklearn + lifelines + scikit-survival
 ```
 
+<Note>
+**There is no `tensorflow` extra**, and `all` does not include TensorFlow — it
+resolves to `pytorch`, `sklearn`, `lifelines` and `scikit-survival`. The
+published extras are exactly: `pytorch`, `sklearn`, `xgboost`, `catboost`,
+`lightgbm`, `lifelines`, `scikit-survival`, `all` and `dev`.
+
+This matters because of how pip behaves: **an unknown extra is a warning, not an
+error, and pip still exits 0**. Asking for one that does not exist therefore
+appears to succeed while installing the core SDK only, and the failure surfaces
+much later as a missing framework.
+</Note>
+
 <Note>
   The umbrella `[boosting]` and `[survival]` extras were removed in 0.10.0, along with `[huggingface]` (the Hugging Face stack now ships inside `[pytorch]`). Use the per-library extras above. `pip` does not fail on an unknown extra — it prints a warning, installs the core SDK only, and exits 0, so the mistake surfaces later as an `ImportError`.
 </Note>