docs: add a little guide on the readme

README.md

@@ -8,7 +8,11 @@ sdk_version: 4.42.0
 app_file: app.py
 pinned: false
 ---
-# ContextqaSV tool
+<p align="center">
+  <img src="assets/contextqasv-logo.png" alt="Imagen" width="200" style="display: inline-block; vertical-align: middle;">
+</p>
+
+# ContextqaSV
 
 HuggingChat tool for querying law documents from El Salvador. This is a forked and customized version of the [RAG community template](https://huggingface.co/spaces/nsarrazin/rag-tool-template/tree/main)
 
@@ -22,5 +26,28 @@ HuggingChat tool for querying law documents from El Salvador. This is a forked a
 * Typed functions.
 * Simplified existing functions.
 
+## How to use?
+
+On the [HuggingChat](https://huggingface.co/chat/) site, follow the next steps:
+
+Click the *Browse more tools* button:
+
+![Brose more tools ss](assets/contextqasv-ss1.png)
+
+Start typing in the search box and click on the **ContextqaSV** tool:
+
+![search and select it ss](assets/contextqasv-ss2.png)
+
+Click on the *Activate* button:
+
+![activate ss](assets/contextqasv-ss3.png)
+
+You're all set! You can now start asking questions about El Salvador's constitution and its traffic regulations:
+
+![start ss](assets/contextqasv-ss4.png)
+
+Note that you can use the [tool's URL](https://hf.co/chat/tools/67a7a96e127c6ef860bce266) to be redirected to HuggingChat and directly show the tool card to activate it (3rd step)
+
 ### Contact
-For inquiries, contributions, or feedback, please contact [me](mailto:[email protected]). Visit my [github account](https://github.com/zaldivards) and check out my other projects :)
+
+For inquiries, contributions, or feedback, please contact [me](mailto:[email protected]). Visit my [github account](https://github.com/zaldivards) and check out my other projects :)

app.py

@@ -54,17 +54,7 @@ def convert(filename: str) -> str:
     ValueError
         If the file type is not supported.
     """
-    plain_text_filetypes = [
-        ".txt",
-        ".csv",
-        ".tsv",
-        ".md",
-        ".yaml",
-        ".toml",
-        ".json",
-        ".json5",
-        ".jsonc",
-    ]
+    plain_text_filetypes = [".txt", ".csv", ".tsv", ".md", ".yaml", ".toml", ".json", ".json5", ".jsonc"]
     # Already a plain text file that wouldn't benefit from pandoc so return the content
     if any(filename.endswith(ft) for ft in plain_text_filetypes):
         with open(filename, "r", encoding="utf-8") as f:
@@ -185,10 +175,11 @@ init()
 gr.Interface(
     predict,
     inputs=[
-        gr.Textbox(label="Query asked about the documents"),
-        gr.Number(label="Number of relevant sources returned (k)", value=default_k),
+        gr.Textbox(label="Question asked about the documents"),
+        gr.Number(label="Number of relevant sources to return (k)", value=default_k),
     ],
     outputs=[gr.Text(label="Relevant chunks")],
-    title="ContextQA tool - El Salvador",
-    description="Forked and customized RAG tool working with law documents from El Salvador",
+    title="ContextqaSV",
+    description="RAG tool enabling questions and answers on legal documents from El Salvador. Legal"
+    " documents supported:\n- Constitución de la república\n- Reglamento de tránsito y seguridad vial",
 ).launch()