jugaad-xbrl

Parse the XBRL filings that Indian listed companies publish on NSE for their quarterly financial results (the SEBI "Integrated Filing – Financials").

• Reads the machine-readable xbrl XML, from a path, raw bytes or a file object.
• Handles all six filing formats: Ind AS, NBFC (Ind AS), Banking, Life Insurance, General Insurance and Other than Bank.
• Joins the facts with the SEBI taxonomy, so you get labelled statements and not just flat elements.
• Offers cross-format fields (net_profit, eps_basic, ...) that mean the same thing for a bank, an insurer and an IT company.

Install

pip install jugaad-xbrl

Python 3.9 or newer. The only runtime dependency is PyYAML.

Usage

from jugaad_xbrl import parse

f = parse("INTEGRATED_FILING_INDAS_..._WEB.xml")   # path, bytes, or binary file object
f.format                  # "indas"
f.reporting_period        # 2026-07-01..2026-09-30

# Cross-format fields. Values are absolute rupees, EPS is rupees per share,
# and None means this format does not file that field.
c = f.canonical()
c.net_profit, c.eps_basic, c.total_assets

# Any raw taxonomy concept for the current reporting period
f.get("ProfitLossForPeriod")

# Labelled statements, with the taxonomy's tree and ordering
stmt = f.statement("100200")          # results; also by role id
for depth, item in stmt.walk():
    print("  " * depth, item.label, item.values)
stmt.rows                             # dimensional rows (segments, related-party transactions, ...)

# Everything in the file, unmodified
f.facts                               # Fact(concept, value, context, unit, decimals)

In-memory use works the same way:

import httpx
f = parse(httpx.get(url).content)

Things to know

• Formats use different element names. For example, net profit is ProfitLossForPeriod (Ind AS, NBFC), ProfitLossForThePeriod (Banking), ProfitLossAfterTaxAndExtraordinaryItems (Life) and ProfitLossAfterTax (General). The canonical layer hides this. jugaad_xbrl/specs/canonical/*.yaml holds the mappings.
• Standalone and consolidated are separate filings. Parse each file on its own. The format is detected from the file.
• Units. Monetary values are always absolute rupees, whatever rounding the filer chose. Ratios are fractions (0.0111 is 1.11%).
• Filers sometimes use the wrong field. Check the ...ExplanatoryTextBlock facts when a number looks odd. The reader keeps them.
• Taxonomy lag. Live filings use the 2026-01-31 taxonomy, but NSE only publishes the 2025-01-31 zips. A few concepts, such as DescriptionOfOther…, are therefore unlabelled. Their facts are still returned.
• The other_than_bank canonical mapping is unverified, as no sample filing was available.

Development

python3 -m venv .venv && .venv/bin/pip install -U pip && .venv/bin/pip install -e ".[dev]"
bash download_specs.sh                        # raw taxonomies and sample filings into ./specs (git-ignored)
.venv/bin/python tools/build_taxonomy.py      # regenerate jugaad_xbrl/specs/taxonomy/*.json
.venv/bin/python -m pytest -q tests

The taxonomy JSON is generated from NSE's native linkbase files and cross-checked in tests against the xlsx that ships in each taxonomy zip. See AGENTS.md for the code layout and nse_integrated_filing_xbrl.md for the filing format reference.

Not affiliated with NSE or SEBI.