config.example.yaml

# config.example.yaml — copy to config.yaml and fill in YOUR OWN values.
#   cp config.example.yaml config.yaml
# config.yaml is git-ignored. Credentials are NEVER in this file — they live in a local
# secret store (see README: LIB_USER / LIB_PASS / *_TDM_KEY via DPAPI on Windows).
#
# This project ships NO institution's access. You supply your own library's endpoints
# (ask your librarian / find them in your library portal) and your own account.

# Your email — required by the Unpaywall API (free, just an identifier). REQUIRED for OA.
unpaywall_email: ""

# Optional: a local sqlite DB (table `papers(doi, oa_pdf_url)`) for an extra OA fallback.
paper_radar_db: ""

# Optional but recommended: your library's holdings table (journal → platform → coverage
# years), built from its A–Z e-journal list. This is what tells you "no entitlement to this
# article" apart from "the route is broken" — see docs/holdings.md for the 6-column schema.
# Blank = holdings.sqlite next to holdings.py, if present; no table at all = checks skipped.
holdings_db: ""

institution:
  # Your library's link resolver (SFX / other OpenURL resolver). Find it in your library
  # portal → e-resources → link resolver. {doi} is substituted at runtime. Leave blank to
  # skip printing a resolver link. Example shape (NOT a real endpoint):
  #   sfx_base: "https://YOUR-SFX-HOST/YOUR-INSTANCE?sid=Claude&id=doi:{doi}"
  sfx_base: ""

  # Your library's off-campus remote-auth gate (the login page for reading full text from
  # home). Only needed for the institutional proxy path (library_session.py).
  #   remote_auth_base: "https://YOUR-REMOTE-AUTH-HOST"
  remote_auth_base: ""

  # Your EZproxy/NetScaler proxy suffix (host[:port]) appended to the dash-rewritten
  # publisher host, e.g. onlinelibrary-wiley-com.<proxy_suffix>. Ask your librarian.
  #   proxy_suffix: "YOUR-PROXY-HOST:PORT"
  proxy_suffix: ""

auth:
  # How library_session.py logs into your remote-auth gate.
  #   form   — a plain login form (EZproxy, Django/NetScaler portals). Works out of the
  #            box once the selectors below match your gate's login page (inspect it in
  #            your browser's devtools).
  #   custom — SSO redirect chains (OpenAthens / Shibboleth): implement login() yourself;
  #            see docs/library-setup.md.
  family: "form"

  # Path of the login form, relative to institution.remote_auth_base.
  #   Django-style gates: "/login/"     EZproxy: "/login"
  login_path: "/login/"

  # CSS selectors for the form fields. Defaults fit a Django-style gate; EZproxy's stock
  # form is: user "input[name='user']", pass "input[name='pass']".
  user_selector: "#id_username"
  pass_selector: "#id_password"
  submit_selector: "form button[type='submit'], form input[type='submit']"

  # Numeric-CAPTCHA support (common on Django-style gates). Leave ALL THREE blank if your
  # gate has no CAPTCHA — the flow degrades to plain fill+submit (the EZproxy case).
  # captcha_image_path may contain {hashkey}, filled from captcha_hashkey_selector's value.
  captcha_selector: ""            # e.g. "#id_captcha_value"
  captcha_hashkey_selector: ""    # e.g. "#hashkey"
  captcha_image_path: ""          # e.g. "/captcha/image/{hashkey}@2/"

  # Session cookies to persist across runs (they are session-only in the browser, so they
  # are saved to your secret store after login and re-injected next run). Map of cookie
  # name -> exact cookie domain — find them in devtools after logging in: the session
  # cookie on the remote-auth host, and the proxy-authorization cookie (often wildcarded
  # on the parent domain). Empty {} = a generic default derived from remote_auth_base.
  persist_cookies: {}
  #   e.g.
  #   persist_cookies:
  #     sessionid: "gate.your-university.edu"
  #     proxy-auth: ".your-university.edu"

rate:
  # Courtesy delay (seconds) between proxy requests. 0 DISABLES it and prints a warning.
  # Systematic/bulk download can get your whole institution's IP blocked — keep this on.
  min_interval_s: 15
  # Your email/handle — sent as a contact User-Agent on direct API routes so publishers
  # can distinguish individual research use from systematic scraping.
  contact: ""