Docker with Elasticsearch

[lipsa7]

from haystack.utils import launch_es
import time

launch_es()
time.sleep(30)

Error:
Tried to start Elasticsearch through Docker but this failed. It is likely that there is already an existing Elasticsearch instance running.

I have installed docker and pulled the image. Can someone please say how to resolve this error?

[anakin87]

Hello, @lipsa7!

Under the hood, launch_es tries to run the following command:

haystack/haystack/utils/doc_store.py

Line 26 in 6e00007

f'docker start {ELASTICSEARCH_CONTAINER_NAME} > /dev/null 2>&1 || docker run -d -p 9200:9200 -e "discovery.type=single-node" --name {ELASTICSEARCH_CONTAINER_NAME} elasticsearch:7.17.6'

In order to debug your problem, please run the following commands using the terminal:

• docker start elasticsearch
• docker run -d -p 9200:9200 -e "discovery.type=single-node" --name elasticsearch elasticsearch:7.17.6

and report here any errors.

[lipsa7]

no
! pip install git+https://github.com/deepset-ai/haystack.git
I used this

[anakin87]

If you want to use Elasticsearch as a document store, you should include the elasticsearch group of dependencies. Please try the aforementioned installation command...

[lipsa7]

import os
from haystack.document_stores import ElasticsearchDocumentStore

host = os.environ.get("ELASTICSEARCH_HOST", "localhost")
document_store = ElasticsearchDocumentStore(host=host, username="", password="", index="document")

It's showing this error 'ConnectionError: Initial connection to Elasticsearch failed. Make sure you run an Elasticsearch instance at [{'host': 'localhost', 'port': 9200}] and that it has finished the initial ramp up (can take > 30s).'

[anakin87]

Now you have some problems with the Elasticsearch container.
Could you paste the output of docker ps -a?

[lipsa7]

CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
34588680f7b9 elasticsearch:7.17.6 "/bin/tini -- /usr/l…" 6 hours ago Up 16 minutes 0.0.0.0:9200->9200/tcp, 9300/tcp elasticsearch

[anakin87]

You can find some useful information about the connection problem in this comment: #1057 (comment)

There are also other related issues:
https://github.com/search?q=repo%3Adeepset-ai%2Fhaystack+ConnectionError+Initial&type=issues

Please let me know if you find a way to make it work...

[lipsa7]

Is there a way to decrease the time please? It took 17 mins for 2 documents (56 MB combined). Thanks a lot!

[anakin87]

I guess it depends on the resources of your machine.
Colab might be faster.

In any case, your two documents seem rather long.

[lipsa7]

It is not giving me the answer I want.

What is HT:LT ratio of Andhra Pradesh East discom?
<Answer {'answer': '13%).', 'type': 'extractive', 'score': 0.9178509712219238, 'context': 'tically higher than the utility \nwith the next highest rate (MP East, at 13%). This \ncould be attributed to unbalanced load, lack of \na protection sys', 'offsets_in_document': [{'start': 235620, 'end': 235625}], 'offsets_in_context': [{'start': 73, 'end': 78}], 'document_ids': ['dc508056555e17df501653fd820cc485'], 'meta': {'name': 'SPI_Electrification_15.pdf'}}>

Text from pdf:
the surveyed utilities
reported an average DT failure rate of 11%, with
three utilities in Gujarat (Central, South, and
North) having low DT failure rates mainly due
to adoption of better maintenance practices.
MP West has the highest DT failure rate, which
at 38% is dramatically higher than the utility
with the next highest rate (MP East, at 13%). This
could be attributed to unbalanced load, lack of
a protection system, and poor workmanship in
the installed DTs.

In its other answers also, it's not giving me the relevant answer to my question. Can you please help in this? Thanks

[anakin87]

This seems a different problem.
Could you open a different discussion?
We can also involve other more knowledgeable people in the discussion.

[lipsa7]

I opened it under discussion 'Not returning the right answers #4941'