Pravah dam storage

The Maharashtra Water Resources Department publishes a storage report every morning on its Pravah dam-safety site. The link serves only the latest report. Pages 1 and 2 summarise storage by region; from page 3 each major dam has a row under its region and district: serial number, name, the date and time of the reading, designed dead, live and gross storage, today’s live and gross storage (Mcum), and today’s live storage as a % of designed live, today and on the same date last year.

The files are in examples/pravah_dams/: recipe.toml and pravah_dams.py. The fixture is pages 5-8 of the 3 October 2026 report, in tests/fixtures/pravah/.

The recipe

name = "Pravah dam storage"

[engines]
ocr = false

[parser]
type = "custom"
function = "pravah_dams.py:parse"
key = ["page", "id", "col"]

# designed gross = dead + live; 0.02 for rounding (Pawana prints 272.11 for 272.13)
[[checks]]
column = "col"
rule = "gross_cap = dead + live_cap"
by = ["page", "id"]
abs_tolerance = 0.02

# today's % = 100 * live / designed live, to 2 decimals
[[checks]]
function = "pravah_dams.py:pct_of_live"

[output]
layout = "table"
rows = ["dam"]
columns = "col"
format = "csv"

The generic rows parser reads nothing here: pdfplumber and PyMuPDF read 98.63 % as two words, so a row ends in %, which is not a value. Serial numbers restart in each district, so id is the dam’s name in lower case with spaces and punctuation dropped. The printed name, region and district are extra columns.

Run it

pdfexorcist extract https://mwrdpravah.in/damsafety/control/pdfLatestReportEng --recipe examples/pravah_dams/recipe.toml -o pravah.csv

Reading Today's-Storage-ReportEng-03-10-2026.pdf from the link.
  pdftotext   1,380 readings  1.5s
  pdfplumber  1,380 readings  46.9s
  pymupdf     1,380 readings  0.8s
  camelot     1,380 readings  35.6s
  pdfium      1,380 readings  0.6s

Agreed        1,380  100.0% of cells
Unresolved        0
Checks      2 rules  all passed

Wrote pravah.csv (table layout, 138 rows).

dam,region,district,sr,date,time,dead,live_cap,gross_cap,live,gross,pct_live,pct_live_last_year
Bawanthadi,Nagpur,Bhandara,1,03/10/2026,08:01 AM,25.56,254.69,280.24,84.00,109.56,32.98,98.87
...
Upper Vaitarna,Nashik,Nashik,15,03/10/2026,08:00 AM,22.65,331.31,353.96,324.18,346.84,97.85,99.83
...
Bhatsa,Kokan,Thane,2,02/10/2026,07:26 AM,34.00,942.10,976.10,929.20,963.20,98.63,98.63
Modaksagar,Kokan,Thane,4,02/10/2026,07:33 AM,76.06,128.92,204.98,101.17,177.23,78.48,99.97
Tansa,Kokan,Thane,5,02/10/2026,07:35 AM,12.08,172.52,184.60,164.83,176.91,95.54,99.55

On the fixture:

pdfexorcist extract tests/fixtures/pravah/pravah_2026-10-03_p5-8.pdf --recipe examples/pravah_dams/recipe.toml

pdfexorcist show tests/fixtures/pravah/pravah_2026-10-03_p5-8.pdf --recipe examples/pravah_dams/recipe.toml --page 4 -o page.png draws what the recipe reads:

Pravah report page with the Thane dams: a row per dam, named in the margin

The parser

parse() in pravah_dams.py, in outline (pseudocode):

def parse(pages):
    for page_no, lines in pages:
        skip the header, down to the column numbers "1 2 3 ... 11"
        for line in lines:
            "Total for ..." or "Grand Total ..."      -> ends the dam above
            a serial, name, date, time, 5 numbers, 2 percentages -> a dam row
            anything else                             -> a loose line, kept for the next row
        before a dam numbered 1, the loose lines are the rest of the previous dam's name,
            then a "... Region" heading, then the district
        before any other dam, they are the rest of the previous dam's name
        each dam: yield {"page", "id", "col": sr ... pct_live_last_year, "value",
                         "dam", "region", "district"} for each value

A dam is held until the next row or total, since the rest of its name or the next district heading can be on the following page.

What is unusual

  • Each dam has its own reading date and time; on the 3 October report they run from 30 September to 3 October.

  • A long name wraps onto a second line: “Wanjarkheda High” and “level Barrage”, with the Parbhani heading right after it. The engines space the second line differently (LB, L B), so the key ignores spaces.

  • The region heading reads “Amravti Region”, the total “Total for Amravati Region”. The region column keeps the heading as printed.

  • Today’s gross storage is not always dead + live: Dimbhe prints 383.78 for 382.05, Thokerwadi Tata 392.35 for 363.70 and Sina kolegaon 39.53 for 61.14. The recipe does not check that column.

tests/test_pravah_dams.py runs the recipe on the fixture and compares every value with the page.