📰 Historical Broadsheet Search (1890–1910)

Cross-Country Comparative Archive • Library of Congress Chronicling America & Pinecone Documents API

📚 46,846 Articles Indexed 🌴 LA Herald 🌁 SF Call 🗽 Evening World ☀️ The Sun 🦅 Chicago Eagle 🌾 Beatrice Express 📅 9 Namespaces
🏛️ Chronological Story Chapters (Click to explore historical narrative):
🔀 Dual-Path Hybrid Query Console Dense Semantic Cosine + Lucene Lexical RRF
Both paths are searched concurrently and fused via Reciprocal Rank Fusion
🧠 Path 1: Semantic Concept Query (Natural Language) Dense Vectors
⚡ Path 2: Full-Text Keyword Query (FTS / BM25) Exact Keywords & Operators
📰 Compare Broadsheets: Selected: 3 of 3 max
Select a story chapter above or customize the dual-path hybrid queries to compare historical broadsheets.
🛠️ Lucene Operators:
📰 Compare Broadsheets Selected: 2 of 3 max
Quick Presets:
🎯 Ranking Engine & Depth Pinecone Integrated Index
Results Depth per Year:
📅 Historical Timeline & Season Window 46,846 Articles across 9 Namespaces
Gilded: Progressive: Season:
🏛️ Curated Inquiries (1-Click Presets Across 6 Publications) Auto-configures query and executes comparison
Enter a query or select a 1-click curated inquiry above to compare broadsheets across the archive.
January
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Historical Curated Dates:
📅 Selected Date: January 01
Explore digitized broadsheets published on this exact calendar day across 1890–1910
Select any calendar date above to explore historical issues published on that exact day across 1890–1910.

Pinecone query_string Grammar & Operator Guide

Capability Syntax Example Archival Use Case & Description
Required Term+aviation +pilotEnsures both terms MUST appear in the article chunk.
Excluded Term-patent -advertisementStrips late-Victorian commercial ads and patent medicine remedies.
Exact Phrase"San Francisco"Matches exact adjacent sequence of tokens in publication order.
Phrase Slop"aviation meet"~4Allows up to 4 words between terms; solves multi-column layout false negatives.
Term Boostingaeroplane^3 "flying machine"^2Multiplies term contribution to score (3x for aeroplane, 2x for flying machine).
Fuzzy / TypoRoosevclt~1, aeroplan~1Tolerates OCR noise & microfilm ink-bleed within Levenshtein distance 1 or 2.
Regex Match/aeronaut.*/Matches regular expression patterns on indexed tokens.
Boolean Logic((a OR b) AND c NOT d)Explicit operator precedence grouping with parentheses.

Nationwide Dataset Architecture (46,846 Articles)

• Archival Broadsheets: 🌴 Los Angeles Herald (8,634 chunks) • 🌁 The San Francisco Call (9,939 chunks) • 🗽 The Evening World (7,829 chunks) • ☀️ The Sun (16,205 chunks) • 🦅 Chicago Eagle (2,439 chunks) • 🌾 The Beatrice Daily Express (1,800 chunks)
• Pinecone Namespaces: Partitioned across 9 year namespaces (1890, 1891, 1892, 1905, 1906, 1907, 1908, 1909, 1910)
• Embedding Model: OpenAI text-embedding-3-small (1,536 dimensions)
• Dynamic Visual CDN: High-resolution Library of Congress IIIF Image API (tile.loc.gov)