invenio.cfg

"""
InvenioRDM settings for CaltechAUTHORS project.

This file was automatically generated by 'invenio-cli init'.

For the full list of settings and their values, see
https://inveniordm.docs.cern.ch/reference/configuration/.
"""

from datetime import datetime
import idutils


def _(x):  # needed to avoid start time failure with lazy strings
    return x

RATELIMIT_ENABLED = True

RATELIMIT_AUTHENTICATED_USER = "50000 per hour;900 per minute"
RATELIMIT_GUEST_USER = "30000 per hour;500 per minute"

# Flask
# =====
# See https://flask.palletsprojects.com/en/1.1.x/config/

# Define the value of the cache control header `max-age` returned by the server when serving
# public files. Files will be cached by the browser for the provided number of seconds.
# See flask documentation for more information:
# https://flask.palletsprojects.com/en/2.1.x/config/#SEND_FILE_MAX_AGE_DEFAULT
SEND_FILE_MAX_AGE_DEFAULT = 300


# Since HAProxy and Nginx route all requests no matter the host header
# provided, the allowed hosts variable is set to localhost. In production it
# should be set to the correct host and it is strongly recommended to only
# route correct hosts to the application.
APP_ALLOWED_HOSTS = ["authors.library.caltech.edu", "0.0.0.0", "localhost", "127.0.0.1"]


# Flask-SQLAlchemy
# ================
# See https://flask-sqlalchemy.palletsprojects.com/en/2.x/config/

# TODO: Set
SQLALCHEMY_DATABASE_URI = (
    "postgresql+psycopg2://caltechauthors:caltechauthors@localhost/caltechauthors"
)


# Invenio-App
# ===========
# See https://invenio-app.readthedocs.io/en/latest/configuration.html

APP_DEFAULT_SECURE_HEADERS = {
    "content_security_policy": {
        "default-src": [
            "'self'",
            "data:",  # for fonts
            "'unsafe-inline'",  # for inline scripts and styles
            "blob:",  # for pdf preview
            # Add your own policies here (e.g. analytics)
        ],
    },
    "content_security_policy_report_only": False,
    "content_security_policy_report_uri": None,
    "force_file_save": False,
    "force_https": True,
    "force_https_permanent": False,
    "frame_options": "sameorigin",
    "frame_options_allow_from": None,
    "session_cookie_http_only": True,
    "session_cookie_secure": True,
    "strict_transport_security": True,
    "strict_transport_security_include_subdomains": True,
    "strict_transport_security_max_age": 31556926,  # One year in seconds
    "strict_transport_security_preload": False,
}


# Flask-Babel
# ===========
# See https://python-babel.github.io/flask-babel/#configuration

# Default locale (language)
BABEL_DEFAULT_LOCALE = "en"
# Default time zone
BABEL_DEFAULT_TIMEZONE = "UTC"


# Invenio-I18N
# ============
# See https://invenio-i18n.readthedocs.io/en/latest/configuration.html

# Other supported languages (do not include BABEL_DEFAULT_LOCALE in list).
I18N_LANGUAGES = [
    # ('de', _('German')),
    # ('tr', _('Turkish')),
]


# Invenio-Theme
# =============
# See https://invenio-theme.readthedocs.io/en/latest/configuration.html

THEME_SITENAME = "CaltechAUTHORS"
# Frontpage title
THEME_FRONTPAGE_TITLE = "CaltechAUTHORS"
# Header logo
THEME_LOGO = "images/CaltechAUTHORS-by-Caltech-Library-blue-with-padding.svg"
THEME_TRACKINGCODE_TEMPLATE = './invenio_theme/trackingcode.html'
# Allow matomo endpoint in the CSP rules
APP_DEFAULT_SECURE_HEADERS['content_security_policy']['default-src'].append(
    'https://cdn.matomo.cloud'
)
APP_DEFAULT_SECURE_HEADERS['content_security_policy']['default-src'].append(
    'https://caltechlibrary.matomo.cloud'
)
# Includes Caltech Library and RAMP
THEME_GOOGLE_SITE_VERIFICATION = ['S9ahhSTJNJrGXH8CLzN9O29lYE5sbaFyCLkdCXHU24U','UqdiAS4yhbCtENKZBNgS0WPiAcMQFShAEjOhMChYmdM']

# Invenio-App-RDM
# ===============
# See https://invenio-app-rdm.readthedocs.io/en/latest/configuration.html

# Instance's theme entrypoint file. Path relative to the ``assets/`` folder.
INSTANCE_THEME_FILE = "./less/theme.less"

# Invenio-Files-Rest
# ==================
FILES_REST_STORAGE_FACTORY = "invenio_s3.s3fs_storage_factory"

# Invenio-S3
# ==========
S3_ENDPOINT_URL = "https://s3.us-west-2.amazonaws.com/"
S3_REGION_NAME = "us-west-2"

# Allow S3 endpoint in the CSP rules
APP_DEFAULT_SECURE_HEADERS["content_security_policy"]["default-src"].append(
    S3_ENDPOINT_URL
)


# Invenio-Records-Resources
# =========================
# See https://github.com/inveniosoftware/invenio-records-resources/blob/master/invenio_records_resources/config.py

SITE_UI_URL = "https://authors.library.caltech.edu"

SITE_API_URL = "https://authors.library.caltech.edu/api"

APP_RDM_DEPOSIT_FORM_DEFAULTS = {
    "publication_date": lambda: datetime.now().strftime("%Y-%m-%d"),
    "rights": [
        {
            "id": "default",
            "title": "No commercial reproduction, distribution, display or performance rights in this work are provided.",
        }
    ],
    "publisher": "Caltech Library",
    "resource_type": {
        "title": {"de": "Zeitschriftenartikel", "en": "Journal Article"},
        "id": "publication-article",
    },
    "languages": [{"id": "eng"}],
}

# See https://github.com/inveniosoftware/invenio-app-rdm/blob/master/invenio_app_rdm/config.py
APP_RDM_DEPOSIT_FORM_AUTOCOMPLETE_NAMES = "search"  # "search_only" or "off"

RDM_CITATION_STYLES = [
    ("chicago-author-date", _("Chicago")),
    ("modern-language-association", _("MLA")),
    ("ieee", _("IEEE")),
    ("apa", _("APA")),
    ("american-chemical-society", _("ACS")),
    ("science", _("Science")),
    ("bibtex", _("Bibtex")),
]

RDM_CITATION_STYLES_DEFAULT = "chicago-author-date"

# Invenio-RDM-Records
# ===================
# See https://inveniordm.docs.cern.ch/customize/dois/
DATACITE_ENABLED = True
DATACITE_USERNAME = "CALTECH.LIBRARY"
DATACITE_PREFIX = "10.7907"
DATACITE_TEST_MODE = False
DATACITE_DATACENTER_SYMBOL = ""

RDM_PERSISTENT_IDENTIFIERS = {
    # DOI automatically removed if DATACITE_ENABLED is False.
    "doi": {
        "providers": ["datacite", "external"],
        "required": False,
        "label": _("DOI"),
        "validator": idutils.is_doi,
        "normalizer": idutils.normalize_doi,
    },
    "oai": {
        "providers": ["oai"],
        "required": True,
        "label": _("OAI"),
    },
}
"""The configured persistent identifiers for records."""

# Authentication - Invenio-Accounts and Invenio-OAuthclient
# =========================================================
# See: https://inveniordm.docs.cern.ch/customize/authentication/

# Invenio-Accounts
# ----------------
# See https://github.com/inveniosoftware/invenio-accounts/blob/master/invenio_accounts/config.py
ACCOUNTS_LOCAL_LOGIN_ENABLED = True  # enable local login
SECURITY_REGISTERABLE = False  # local login: allow users to register
SECURITY_RECOVERABLE = True  # local login: allow users to reset the password
SECURITY_CHANGEABLE = True  # local login: allow users to change psw
SECURITY_CONFIRMABLE = True  # local login: users can confirm e-mail address
SECURITY_LOGIN_WITHOUT_CONFIRMATION = (
    False  # require users to confirm email before being able to login
)
ACCOUNTS_DEFAULT_EMAIL_VISIBILITY= "public"
ACCOUNTS_DEFAULT_USER_VISIBILITY = "public"

# Invenio-OAuthclient
# -------------------
# See https://github.com/inveniosoftware/invenio-oauthclient/blob/master/invenio_oauthclient/config.py

OAUTHCLIENT_REMOTE_APPS = {}  # configure external login providers

from invenio_oauthclient.views.client import auto_redirect_login

ACCOUNTS_LOGIN_VIEW_FUNCTION = (
    auto_redirect_login  # autoredirect to external login if enabled
)
OAUTHCLIENT_AUTO_REDIRECT_TO_EXTERNAL_LOGIN = False  # autoredirect to external login

# Invenio-UserProfiles
# --------------------
USERPROFILES_READ_ONLY = (
    False  # allow users to change profile info (name, email, etc...)
)

# OAI-PMH
# =======
# See https://github.com/inveniosoftware/invenio-oaiserver/blob/master/invenio_oaiserver/config.py

OAISERVER_ID_PREFIX = "authors.library.caltech.edu"
"""The prefix that will be applied to the generated OAI-PMH ids."""

# Invenio-Search
# --------------

SEARCH_INDEX_PREFIX = "caltechauthors-"
from invenio_rdm_records.config import RDM_SORT_OPTIONS
RDM_SORT_OPTIONS.update({
    "publication": dict(
        title=_('Publication date'),
        fields=['-metadata.publication_date'],
    ),
    "newest": dict(
        title=_("Recently added"),
        fields=["-created"],
    ),
})
RDM_SEARCH = {
    "facets": ["resource_type", "subject", "access_status", "file_type" ],
    "sort": [
        "bestmatch",
        "newest",
	"publication",
        "version",
        "mostviewed",
        "mostdownloaded",
    ],
}


SECURITY_EMAIL_SENDER = "repositories@library.caltech.edu"
MAIL_SERVER = "smtp.gmail.com"
MAIL_USERNAME = "repositories@library.caltech.edu"
MAIL_USE_SSL = True
MAIL_PORT = 465
SECURITY_EMAIL_SUBJECT_PASSWORD_RESET = "Reset your password for CaltechAUTHORS"
"""Email subject for account registration emails."""

APP_RDM_ADMIN_EMAIL_RECIPIENT = "dld-dev@library.caltech.edu"

MAIL_SUPPRESS_SEND = False

# Customize identifiers
# --------------------


def always_valid(identifier):
    """Gives every identifier as valid."""
    return True


RDM_RECORDS_IDENTIFIERS_SCHEMES = {
    "ark": {"label": _("ARK"), "validator": idutils.is_ark, "datacite": "ARK"},
    "arxiv": {"label": _("arXiv"), "validator": idutils.is_arxiv, "datacite": "arXiv"},
    "bibcode": {
        "label": _("Bibcode"),
        "validator": idutils.is_ads,
        "datacite": "bibcode",
    },
    "doi": {"label": _("DOI"), "validator": idutils.is_doi, "datacite": "DOI"},
    "ean13": {"label": _("EAN13"), "validator": idutils.is_ean13, "datacite": "EAN13"},
    "eissn": {"label": _("EISSN"), "validator": idutils.is_issn, "datacite": "EISSN"},
    "handle": {
        "label": _("Handle"),
        "validator": idutils.is_handle,
        "datacite": "Handle",
    },
    "igsn": {"label": _("IGSN"), "validator": always_valid, "datacite": "IGSN"},
    "isbn": {"label": _("ISBN"), "validator": idutils.is_isbn, "datacite": "ISBN"},
    "issn": {"label": _("ISSN"), "validator": idutils.is_issn, "datacite": "ISSN"},
    "istc": {"label": _("ISTC"), "validator": idutils.is_istc, "datacite": "ISTC"},
    "lissn": {"label": _("LISSN"), "validator": idutils.is_issn, "datacite": "LISSN"},
    "lsid": {"label": _("LSID"), "validator": idutils.is_lsid, "datacite": "LSID"},
    "pmcid": {"label": _("PMCID"), "validator": idutils.is_pmcid, "datacite": "PMID"},
    "pmid": {"label": _("PMID"), "validator": idutils.is_pmid, "datacite": "PMID"},
    "purl": {"label": _("PURL"), "validator": idutils.is_purl, "datacite": "PURL"},
    "upc": {"label": _("UPC"), "validator": always_valid, "datacite": "UPC"},
    "url": {"label": _("URL"), "validator": idutils.is_url, "datacite": "URL"},
    "urn": {"label": _("URN"), "validator": idutils.is_urn, "datacite": "URN"},
    "w3id": {"label": _("W3ID"), "validator": always_valid, "datacite": "w3id"},
    "eprintid": {
        "label": _("Eprint ID"),
        "validator": always_valid,
        "datacite": "Eprint ID",
    },
    "resolverid": {
        "label": _("Resolver ID"),
        "validator": always_valid,
        "datacite": "Resolver ID",
    },
    "patentnumber": {
        "label": _("Patent Number"),
        "validator": always_valid,
        "datacite": "Patent Number",
    },
    "patentnumber": {
        "label": _("Patent Number"),
        "validator": always_valid,
        "datacite": "Patent Number",
    },
    "ostiid": {
        "label": _("OSTI ID"),
        "validator": idutils.is_pmid,
        "datacite": "OSTI ID",
    },
}
"""These are used for main, alternate and related identifiers."""

RDM_RECORDS_PERSONORG_SCHEMES = {
    "orcid": {"label": _("ORCID"), "validator": idutils.is_orcid, "datacite": "ORCID"},
    "isni": {"label": _("ISNI"), "validator": idutils.is_isni, "datacite": "ISNI"},
    "gnd": {"label": _("GND"), "validator": idutils.is_gnd, "datacite": "GND"},
    "ror": {"label": _("ROR"), "validator": idutils.is_ror, "datacite": "ROR"},
    "clpid": {
        "label": _("Caltech Library Person ID"),
        "validator": idutils.is_clpid,
        "datacite": "CLPID",
    },
}

VOCABULARIES_NAMES_SCHEMES = {
    "orcid": {"label": _("ORCID"), "validator": idutils.is_orcid, "datacite": "ORCID"},
    "isni": {"label": _("ISNI"), "validator": idutils.is_isni, "datacite": "ISNI"},
    "gnd": {"label": _("GND"), "validator": idutils.is_gnd, "datacite": "GND"},
    "ror": {"label": _("ROR"), "validator": idutils.is_ror, "datacite": "ROR"},
    "clpid": {
        "label": _("Caltech Library Person ID"),
        "validator": idutils.is_clpid,
        "datacite": "CLPID",
    },
}

# Enable featured communities
COMMUNITIES_ADMINISTRATION_DISABLED = False

# Basic homepage customization
THEME_SHOW_FRONTPAGE_INTRO_SECTION = False

from invenio_i18n import lazy_gettext as _
from invenio_rdm_records.contrib.imprint import (
    IMPRINT_CUSTOM_FIELDS,
    IMPRINT_CUSTOM_FIELDS_UI,
    IMPRINT_NAMESPACE,
)
from invenio_rdm_records.contrib.journal import (
    JOURNAL_CUSTOM_FIELDS,
    JOURNAL_CUSTOM_FIELDS_UI,
    JOURNAL_NAMESPACE,
)
from invenio_rdm_records.contrib.thesis import (
    THESIS_CUSTOM_FIELDS,
    THESIS_CUSTOM_FIELDS_UI,
    THESIS_NAMESPACE,
)
from invenio_rdm_records.contrib.meeting import (
    MEETING_CUSTOM_FIELDS,
    MEETING_CUSTOM_FIELDS_UI,
    MEETING_NAMESPACE,
)
from invenio_vocabularies.services.custom_fields import VocabularyCF
from invenio_records_resources.services.custom_fields import TextCF

RDM_NAMESPACES = {
    **JOURNAL_NAMESPACE,
    **IMPRINT_NAMESPACE,
    **THESIS_NAMESPACE,
    **MEETING_NAMESPACE,
    "caltech": "https://library.caltech.edu",
}

# Combine custom fields
RDM_CUSTOM_FIELDS = [
    # journal
    *JOURNAL_CUSTOM_FIELDS,
    # meeting
    *MEETING_CUSTOM_FIELDS,
    # imprint
    *IMPRINT_CUSTOM_FIELDS,
    # thesis
    *THESIS_CUSTOM_FIELDS,
    # caltech groups
    VocabularyCF(
        name="caltech:groups",
        dump_options=True,
        vocabulary_id="caltechgroups",
        multiple=True,
    ),
    # Custom series and volume/issue
    TextCF(name="caltech:series"),
    TextCF(name="caltech:series_number"),
    TextCF(name="caltech:other_num_name"),
    TextCF(name="caltech:other_num_id"),
    # caltech journal list
    VocabularyCF(
        name="caltech:journals", dump_options=False, vocabulary_id="caltechjournals"
    ),
    VocabularyCF(
        name="caltech:publication_status", dump_options=True, vocabulary_id="publicationstatus",multiple=True
    ),
]

# Hide "meeting" section in the landing page
MEETING_CUSTOM_FIELDS_UI["hide_from_landing_page"] = True

# Create a "Publishing Information" section in the deposit form and add multiple fields
RDM_CUSTOM_FIELDS_UI = [
    {
        "section": _("Publishing information"),
        "hide_from_landing_page": True,
        "fields": [
            # journal
            *JOURNAL_CUSTOM_FIELDS_UI["fields"],
            # imprint
            *IMPRINT_CUSTOM_FIELDS_UI["fields"],
            # thesis
            *THESIS_CUSTOM_FIELDS_UI["fields"],
        ],
    },
    {
        "section": _("Caltech Custom Metadata"),
        "fields": [
            dict(
                field="caltech:groups",
                ui_widget="Dropdown",
                props=dict(
                    label="Caltech groups",
                    placeholder="Search for a Caltech group by name or identifier...",
                    icon="users",
                    description="The group or groups associated with record.",
                    autocompleteFrom="/api/vocabularies/caltechgroups",
                    search=True,
                    multiple=True,
                    clearable=True,
                ),
            ),
            dict(
                field="caltech:series",
                ui_widget="Input",
                props=dict(
                    label="Series Name",
                    placeholder="Technical Report Series",
                    icon="archive",
                    description="The name of a series.",
                ),
            ),
            dict(
                field="caltech:series_number",
                ui_widget="Input",
                props=dict(
                    label="Series Volume or Issue Number",
                    placeholder="5",
                    description="The volume or issue number in a series.",
                ),
            ),
            dict(
                field="caltech:other_num_name",
                ui_widget="Input",
                props=dict(
                    label="Other Numbering System Name",
                    placeholder="LIGO",
                    icon="calculator",
                    description="A author-assigned or other numbering system used for the item.",
                ),
            ),
            dict(
                field="caltech:other_num_id",
                ui_widget="Input",
                props=dict(
                    label="Other Numbering System Identifier",
                    placeholder="L123-456",
                    description="The identifier of this item in an author-assigned or other numbering system.",
                ),
            ),
            dict(
                field="caltech:journals",
                ui_widget="AutocompleteDropdown",
                props=dict(
                    label="Journal name",
                    placeholder="Search for a Journal on the CaltechAUTHORS journal list...",
                    icon="address book",
                    description="If you select a journal listed here, you do not need to complete the Journal, ISSN, and Publisher fields.",
                    autocompleteFrom="/api/vocabularies/caltechjournals",
                    search=True,
                    multiple=False,
                    clearable=True,
                ),
            ),
	   dict(
                field="caltech:publication_status",
                ui_widget="Dropdown",
                props=dict(
                    label="Publication Status",
                    placeholder="Add publication status tags...",
                    icon="address book",
                    autocompleteFrom="/api/vocabularies/publicationstatus",
                    search=True,
                    multiple=True,
                    clearable=True,
                ),
            ),
        ],
    },
    # meeting
    MEETING_CUSTOM_FIELDS_UI,
]

import sentry_sdk
from flask import Flask
from sentry_sdk.integrations.flask import FlaskIntegration

sentry_sdk.init(
    dsn="https://62229a4fbf83a2bd327820cd1c9fffe2@o4505830343245824.ingest.sentry.io/4505830344818688",
    integrations=[FlaskIntegration()],
    # Set traces_sample_rate to 1.0 to capture 100%
    # of transactions for performance monitoring.
    # We recommend adjusting this value in production.
    traces_sample_rate=0.01,
    # Set profiles_sample_rate to 1.0 to profile 100%
    # of sampled transactions.
    # We recommend adjusting this value in production.
    profiles_sample_rate=0.01,
)

OAISERVER_RESUMPTION_TOKEN_EXPIRE_TIME = 1 * 120
OAISERVER_PAGE_SIZE = 100
OAISERVER_ADMIN_EMAILS = [
    "coda@library.caltech.edu",
]