Demo UI add env vars & other small fixes (deepset-ai#1828)

* Add more env vars to the streamlit ui

* Add some more questions to the random ones

* Relax a statuscode check and rename env vars

* Make query error message more descriptive

* Add log message

* Align docker-compose with and without GPU

* Typo in pipeline filename

* Remove prefix from var in docker_compose

* Align docker-compose.yml and add small sleep to the initialized poller to prevent spamming

* Fix the name of the dockerfile used to build the GPU image

docker-compose-gpu.yml

@@ -4,7 +4,7 @@ services:
   haystack-api:
     build:
       context: .
-      dockerfile: Dockerfile
+      dockerfile: Dockerfile-GPU
     image: "deepset/haystack-gpu:latest"
     # in recent docker-compose version you can enable GPU resources. Make sure to fulfill the prerequisites listed here: https://docs.docker.com/compose/gpu-support/
     deploy:
@@ -52,5 +52,11 @@ services:
     environment:
       - API_ENDPOINT=http://haystack-api:8000
       - EVAL_FILE=eval_labels_example.csv
-      - HAYSTACK_UI_DISABLE_FILE_UPLOAD   # docker-compose run -e HAYSTACK_UI_DISABLE_FILE_UPLOAD=1 ...
+      # The value fot the following variables will be read from the host, if present.
+      # They can also be temporarily set for docker-compose, for example:
+      # DISABLE_FILE_UPLOAD=1 DEFAULT_DOCS_FROM_RETRIEVER=5 docker-compose up
+      - DISABLE_FILE_UPLOAD
+      - DEFAULT_QUESTION_AT_STARTUP
+      - DEFAULT_DOCS_FROM_RETRIEVER
+      - DEFAULT_NUMBER_OF_ANSWERS
     command: "/bin/bash -c 'sleep 15 && streamlit run webapp.py'"

docker-compose.yml

@@ -14,9 +14,11 @@ services:
     environment:
       # See rest_api/pipelines.yaml for configurations of Search & Indexing Pipeline.
       - DOCUMENTSTORE_PARAMS_HOST=elasticsearch
+      - PIPELINE_YAML_PATH=/home/user/rest_api/pipeline/pipelines.yaml
+      - CONCURRENT_REQUEST_PER_WORKER
     depends_on:
       - elasticsearch
-    command: "/bin/bash -c 'sleep 10 && gunicorn rest_api.application:app -b 0.0.0.0 -k uvicorn.workers.UvicornWorker --workers 1 --timeout 180'"
+    command: "/bin/bash -c 'sleep 10 && gunicorn rest_api.application:app -b 0.0.0.0 -k uvicorn.workers.UvicornWorker --workers 8 --timeout 180'"
   elasticsearch:
     # This will start an empty elasticsearch instance (so you have to add your documents yourself)
     #image: "elasticsearch:7.9.2"
@@ -39,4 +41,11 @@ services:
     environment:
       - API_ENDPOINT=http://haystack-api:8000
       - EVAL_FILE=eval_labels_example.csv
+      # The value fot the following variables will be read from the host, if present.
+      # They can also be temporarily set for docker-compose, for example:
+      # DISABLE_FILE_UPLOAD=1 DEFAULT_DOCS_FROM_RETRIEVER=5 docker-compose up
+      - DISABLE_FILE_UPLOAD
+      - DEFAULT_QUESTION_AT_STARTUP
+      - DEFAULT_DOCS_FROM_RETRIEVER
+      - DEFAULT_NUMBER_OF_ANSWERS
     command: "/bin/bash -c 'sleep 15 && streamlit run webapp.py'"

rest_api/controller/search.py

@@ -29,6 +29,7 @@
 logging.info(f"Loaded pipeline nodes: {PIPELINE.graph.nodes.keys()}")
 
 concurrency_limiter = RequestLimiter(CONCURRENT_REQUEST_PER_WORKER)
+logging.info("Concurrent requests per worker: {CONCURRENT_REQUEST_PER_WORKER}")
 
 
 @router.get("/initialized")

ui/eval_labels_example.csv

@@ -1,4 +1,13 @@
 "Question Text";"Answer"
-"What's the capital of France?";"Paris"
-"What's the tallest mountain in Kenya?";"Mount Kenya"
-"What's the climate of Beijing?";"humid continental"
+"What is the capital of the Netherlands?";"Amsterdam"
+"What's the tallest mountain in Africa?";"Mount Kilimanjaro"
+"What's the climate of Beijing?";"monsoon-influenced humid continental"
+"What's the longest river of Europe?";"The Volga"
+"What's the deepest lake in the world?";"Lake Bajkal"
+"How many people live in the capital of the US?";"689,545"
+"Which Chinese ciy is the largest?";"Shanghai"
+"What's the type of government of the UK?";"unitary parliamentary democracy and constitutional monarchy"
+"What currency is used in Hungary?";"Hungarian forint"
+"In which city is the Louvre?";"Paris"
+"Which language is spoken in the Fiji?";"Fiji Hindi"
+"Who is the current king of Spain?";"Felipe VI"

ui/utils.py

@@ -3,6 +3,7 @@
 import os
 import logging
 import requests
+from time import sleep
 from uuid import uuid4
 import streamlit as st
 
@@ -25,6 +26,7 @@ def haystack_is_ready():
             return True
     except Exception as e:
         logging.exception(e)
+        sleep(1)  # To avoid spamming a non-existing endpoint at startup
     return False
 
 
@@ -48,8 +50,8 @@ def query(query, filters={}, top_k_reader=5, top_k_retriever=5) -> Tuple[List[Di
     req = {"query": query, "params": params}
     response_raw = requests.post(url, json=req)
 
-    if response_raw.status_code >= 400:
-        raise Exception(f"{response_raw}")
+    if response_raw.status_code >= 400 and response_raw.status_code != 503:
+        raise Exception(f"{vars(response_raw)}")
 
     response = response_raw.json()
     if "errors" in response:

ui/webapp.py

@@ -17,13 +17,17 @@
 
 
 # Adjust to a question that you would like users to see in the search bar when they load the UI:
-DEFAULT_QUESTION_AT_STARTUP = "What's the capital of France?"
+DEFAULT_QUESTION_AT_STARTUP = os.getenv("DEFAULT_QUESTION_AT_STARTUP", "What's the capital of France?")
+
+# Sliders
+DEFAULT_DOCS_FROM_RETRIEVER = int(os.getenv("DEFAULT_DOCS_FROM_RETRIEVER", 3))
+DEFAULT_NUMBER_OF_ANSWERS = int(os.getenv("DEFAULT_NUMBER_OF_ANSWERS", 3))
 
 # Labels for the evaluation
 EVAL_LABELS = os.getenv("EVAL_FILE", Path(__file__).parent / "eval_labels_example.csv")
 
 # Whether the file upload should be enabled or not
-DISABLE_FILE_UPLOAD = os.getenv("HAYSTACK_UI_DISABLE_FILE_UPLOAD")
+DISABLE_FILE_UPLOAD = bool(os.getenv("DISABLE_FILE_UPLOAD"))
 
 
 def main():
@@ -54,8 +58,20 @@ def reset_results(*args):
 
     # Sidebar
     st.sidebar.header("Options")
-    top_k_reader = st.sidebar.slider("Max. number of answers", min_value=1, max_value=10, value=3, step=1, on_change=reset_results)
-    top_k_retriever = st.sidebar.slider("Max. number of documents from retriever", min_value=1, max_value=10, value=3, step=1, on_change=reset_results)
+    top_k_reader = st.sidebar.slider(
+        "Max. number of answers", 
+        min_value=1, 
+        max_value=10, 
+        value=DEFAULT_NUMBER_OF_ANSWERS, 
+        step=1, 
+        on_change=reset_results)
+    top_k_retriever = st.sidebar.slider(
+        "Max. number of documents from retriever", 
+        min_value=1, 
+        max_value=10, 
+        value=DEFAULT_DOCS_FROM_RETRIEVER, 
+        step=1, 
+        on_change=reset_results)
     eval_mode = st.sidebar.checkbox("Evaluation mode")
     debug = st.sidebar.checkbox("Show debug info")
 
@@ -158,7 +174,7 @@ def reset_results(*args):
                 return
             except Exception as e:
                 logging.exception(e)
-                if "The server is busy processing requests" in str(e):
+                if "The server is busy processing requests" in str(e) or "503" in str(e):
                     st.error("🧑‍🌾    All our workers are busy! Try again later.")
                 else:
                     st.error("🐞    An error occurred during the request.")