# FTP mirror recovery — 2026-09-03

Read-only FTP mirror of shodh.net (`scripts/legacy_ftp_mirror.py`), run to recover
files that the August web crawl could not reach — principally any Vol.15/16
print-copy PDFs, which the crawl had only seen behind cPanel-session links.

Mirror output (gitignored, personal/site data): `legacy-import/backup/files/`

| | |
|---|---|
| Files downloaded | 839 |
| Total size | 284.6 MB |
| Folders mirrored | `phocadownload`, `media`, `ugc_pdf`, `images`, `templates/shodh/images`, `shodh-forms` |
| Write operations issued | none — LIST/RETR only |

## Vol.15 and Vol.16 — the headline result

Both volumes exist as Joomla **menu entries** (`jos_menu` rows 189/192 → category 63,
190/193 → category 64) and as **empty Phoca categories**. Neither has a single file
record in `jos_phocadownload`.

- **Vol.15** — `phocadownload/vol-15-issue-1-2/` is an **empty directory** on the
  server. Nothing was ever uploaded. There is nothing to recover, from any source.
- **Vol.16** — two PDFs sit in `phocadownload/vol-16-issue-1-2/` but were **never
  registered as downloads** in Joomla:
  - `1_Vol. 16_अनुज प्रताप सिंह_नाथ-साहित्य और हठयोग.pdf` (152 KB)
  - `2_Vol. 16_Dr. Arvind Kumar Singh_An Analytical Study of Copyright Act Violations in India.pdf` (171 KB)

Because they have no Phoca record, the legacy site never listed or served them —
visitors clicking "Vol. 16, Issue 1 & 2" saw an empty category page. They are
uploaded-but-unpublished files.

**These two files have deliberately NOT been imported.** Publishing them would put
content on the journal's website that has never been public, which is an editorial
decision for the chief editor, not a migration step. The files are preserved in the
mirror and can be imported in minutes if approved.

## What the mirror did recover

Comparing all 572 mirrored PDFs against the 466 already imported (by SHA-256):

- **78** mirrored PDFs were not in the archive.
- **69** of those are files Joomla never served: duplicate renderings under a second
  filename (most of Vol.8 Issue 1 and Vol.3 Issue 1), whole-issue compilations
  (`*_Complete_File.pdf`, `All_Artical.pdf`), and stray uploads. Not gaps.
- **9** had live Phoca records. Two were already imported under a different byte
  variant (legacy ids 13 and 259). Four belong to news/announcement categories
  (38, 53), not to journal issues — impact-factor certificate, admission notice,
  workshop news, seminar file.
- **2 were genuine archive gaps and have been imported** (see below).

## Imported

Manifest: `legacy-import/manifest-recovered.json`. Applied through the existing
idempotent admin import (`POST /admin/archive/import`), keyed on legacy file id, so
re-running is safe.

| Legacy id | Issue | Title | Legacy downloads |
|---|---|---|---|
| 263 | Vol.5, Issue 1 | मालिक मुहम्मद जायसी कृत 'अखरावट' में रहस्यवाद — धर्मेन्द्र कुमार | 1,583 |
| 327 | Vol.6, Issue 2 | भारत के उत्तरी सीमा की विक्षोभजनित चिन्ताएँ एवं नेपाल (अग्रलेख) — योगेन्द्र प्रताप सिंह | 1,081 |

Both were published on the legacy site with real download histories; the crawl had
missed them because their links were broken there. Archive article total: 477 → 479.

Verified after import: both resolve through the legacy file-id route and download as
valid PDFs (218,111 and 42,102 bytes, `%PDF` header, `application/pdf`).

## Also recovered, not yet used

The mirror picked up site assets the crawl never had, held in case they are wanted:

- `media/` — membership forms (Hindi and English), certification-of-original-work
  form, and `hindifont.zip` (the legacy Walkman-Chanakya font bundle)
- `ugc_pdf/` — UGC journal list and both gazette notifications, including the English
  gazette that the crawl could not download
- `templates/shodh/images/` and `images/shodh/` — original masthead, banner and icon
  artwork
- `shodh-forms/` — the PHP source of the legacy registry application, which documents
  the original registry field semantics

## Open items

- Vol.16's two files await an editorial decision from Dr. Y.P. Singh: publish or leave
  unpublished, matching the legacy site.
- Vol.15 has no content anywhere. If the issue was in fact published in print, the
  PDFs would have to come from the editor's own records.