[Core] Consolidate prompt arguments to LLM engines (#4328)

Co-authored-by: Roger Wang <ywang@roblox.com>

docs/source/dev/offline_inference/llm.rst

@@ -1,5 +1,5 @@
 LLM Class
-==========
+=========
 
 .. autoclass:: vllm.LLM
     :members:

docs/source/dev/offline_inference/llm_inputs.rst

@@ -0,0 +1,14 @@
+LLM Inputs
+==========
+
+.. autodata:: vllm.inputs.PromptStrictInputs
+
+.. autoclass:: vllm.inputs.TextPrompt
+    :show-inheritance:
+    :members:
+    :member-order: bysource
+
+.. autoclass:: vllm.inputs.TokensPrompt
+    :show-inheritance:
+    :members:
+    :member-order: bysource

docs/source/dev/offline_inference/offline_index.rst

@@ -0,0 +1,8 @@
+Offline Inference
+=================================
+
+.. toctree::
+   :maxdepth: 1
+
+   llm
+   llm_inputs

docs/source/index.rst

@@ -68,13 +68,6 @@ Documentation
    getting_started/quickstart
    getting_started/examples/examples_index
 
-.. toctree::
-   :maxdepth: 1
-   :caption: Offline Inference
-
-   offline_inference/llm
-   offline_inference/sampling_params
-
 .. toctree::
    :maxdepth: 1
    :caption: Serving
@@ -108,7 +101,9 @@ Documentation
 .. toctree::
    :maxdepth: 2
    :caption: Developer Documentation
-
+   
+   dev/sampling_params
+   dev/offline_inference/offline_index
    dev/engine/engine_index
    dev/kernel/paged_attention
    dev/dockerfile/dockerfile

docs/source/serving/openai_compatible_server.md

@@ -48,7 +48,7 @@ completion = client.chat.completions.create(
 ```
 
 ### Extra Parameters for Chat API
-The following [sampling parameters (click through to see documentation)](../offline_inference/sampling_params.rst) are supported.
+The following [sampling parameters (click through to see documentation)](../dev/sampling_params.rst) are supported.
 
 ```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
 :language: python
@@ -65,7 +65,7 @@ The following extra parameters are supported:
 ```
 
 ### Extra Parameters for Completions API
-The following [sampling parameters (click through to see documentation)](../offline_inference/sampling_params.rst) are supported.
+The following [sampling parameters (click through to see documentation)](../dev/sampling_params.rst) are supported.
 
 ```{literalinclude} ../../../vllm/entrypoints/openai/protocol.py
 :language: python