Solr Integration in Python — Core Concepts

Apache Solr is an open-source search platform built on Apache Lucene. It provides full-text search, faceted navigation, highlighting, and spell-checking. Python connects to Solr through HTTP-based clients, most commonly pysolr.

Connecting with pysolr

import pysolr

solr = pysolr.Solr('http://localhost:8983/solr/my_collection', always_commit=True, timeout=10)

# Health check
solr.ping()

For SolrCloud (distributed mode), connect through ZooKeeper:

zookeeper = pysolr.ZooKeeper("zk1:2181,zk2:2181,zk3:2181")
solr = pysolr.SolrCloud(zookeeper, "my_collection")

Indexing documents

Documents are Python dictionaries. Each needs a unique id field.

docs = [
    {"id": "1", "title": "Python Async Guide", "content": "Asyncio enables...", "tags": ["python", "async"]},
    {"id": "2", "title": "Django REST Patterns", "content": "Building APIs with...", "tags": ["python", "django"]},
]

solr.add(docs)

Solr automatically detects field types if using schemaless mode, but production systems should define explicit schemas.

Searching

Basic query

results = solr.search("async programming", rows=10)

print(f"Found {results.hits} results")
for doc in results:
    print(f"{doc['title']}")

Field-specific queries

Solr uses Lucene query syntax:

# Search title field specifically
results = solr.search("title:python AND content:testing")

# Phrase search
results = solr.search('"machine learning"')

# Range queries
results = solr.search("published:[2026-01-01T00:00:00Z TO *]")

Boosting

# Title matches are 3x more important
results = solr.search("title:python^3 OR content:python")

Faceted search

Facets let users drill down into results by category — like filters on an e-commerce site.

results = solr.search("python", **{
    'facet': 'on',
    'facet.field': ['tags', 'category'],
    'facet.mincount': 1,
    'rows': 10
})

# Access facet counts
for field, counts in results.facets['facet_fields'].items():
    print(f"\n{field}:")
    # counts is a flat list: [value, count, value, count, ...]
    for i in range(0, len(counts), 2):
        print(f"  {counts[i]}: {counts[i+1]}")

Highlighting

Solr can return snippets with matching terms highlighted:

results = solr.search("async programming", **{
    'hl': 'true',
    'hl.fl': 'content',
    'hl.simple.pre': '<mark>',
    'hl.simple.post': '</mark>',
    'hl.fragsize': 200
})

for doc in results:
    doc_id = doc['id']
    highlights = results.highlighting.get(doc_id, {})
    print(highlights.get('content', ['No highlight'])[0])

Spell checking and suggestions

results = solr.search("pythn programing", **{
    'spellcheck': 'true',
    'spellcheck.collate': 'true',
    'spellcheck.count': 5
})
# Returns suggested corrections

When to choose Solr

Scenario	Solr fits well
Enterprise search with complex faceting	Yes — faceting is a first-class feature
Existing Java/Lucene infrastructure	Yes — natural fit
Need for fine-grained XML configuration	Yes — Solr’s config is very customizable
Real-time analytics and dashboards	Less ideal — Elasticsearch has stronger analytics tooling
Simple setup, minimal ops	Less ideal — Solr requires more upfront configuration

Common misconception

People think Solr is outdated because Elasticsearch gets more attention. Solr remains actively maintained and is used at massive scale. Bloomberg, Netflix, and Apple run Solr in production. The choice between them is often about ecosystem fit, not capability.

One thing to remember: Solr gives Python powerful search with facets, highlighting, and spell-check — it thrives in environments that value configuration control and proven stability.