💾 Archived View for radia.bortzmeyer.org › software › agunua captured on 2024-07-09 at 05:40:59. Gemini links have been rewritten to link to archived content

-=-=-=-=-=-=-

Agunua

Agunua is a Python library for the development of Gemini clients.

Installation

You need Python 3, PySocks, PyOpenSSL and netaddr. You can install the dependencies with pip `pip3 install agunua`.

PyOpenSSL

PySocks

netaddr

Usage

u = agunua.GeminiUri(url)
print(u)

Parameters in the `GeminiUri()` constructor (you can find their

default values at the beginning of the file `Agunua/__init__.py`):

`url`: the URL to load
`insecure`: accept invalid certificates (signed by unknown CA, for instance)
`tofu`: performs TOFU (Trust On First Use) validation
`accept_expired`: accept expired certificates (`insecure = True` is not sufficient for that)
`get_content`: retrieve the actual resource (default is to get metadata only)
`parse_content`: if it is gemtext (text/gemini), parse and extract links
`maxlines`: if it is text, maximum number of lines retrieved. Set to None if you don't want a limit
`maxsize`: maximum size in bytes to retrieve. Set to None if you don't want a limit
`binary`: (automatic if the content is text). Retrieve as binary content, don't play with end-of-lines
`follow-redirect`: automatically follow Gemini redirections
`redirect_depth`: maximum number of redirections followed
`iri`: handle IRI (URI in Unicode)
`force_ipv4`: use the IPv4 protocol only
`force_ipv6`: use the IPv6 protocol only
`send_sni`: send the TLS Server Name Indication
`connect_to`: use the host name in the URI for the Gemini request but connect only to this host (name or address). Useful when the host is multihomed.
`clientcert`: the filename of a client certificate that will be sent to the server.
`clientkey`: the filename of the private key of the above certificate.
`use_socks`: use a SOCKS5 proxy (for instance for `.onion` capsules). The value must be a tuple (socks proxy name, socks proxy port).

If the URL is invalid (wrong syntax), you won't get a `GeminiUri` object. If you get one, it does not mean the resource has been retrieved successfully. See the attribute `network_success` for that, and then the attribute `status_code` (that you have to interpret yourself, in most cases).

Attributes of `GeminiUri` objects (not all of them will always be present; for instance, if you did not ask to get content, you won't have an attribute `size`; if the status code is not 20 - OK - you won't get a mediatype; etc):

`network_success`: resource was retrieved successfully
`status_code`: if retrieved successfully, the Gemini two-digit status code
`error`: if `network_success` is false, this is the reason
`ip_address`: IP address used for the retrieval (except is SOCKS was used)
`meta`: the `meta` field of the Gemini protocol. It depends on the status code. Read the Gemini specification for detail.
`binary`: if you asked for binary access, it will be True. If you asked for text access (binary=False in the constructor) and asked to ge the content (get_content=True), it will be set to False if decoding went well and True if the decoding failed, for instance because the file did not match the announced "charset".
`links`: an array of the links found in the document (if you've set `parse_content`)
`payload`: the content
`size`: the size of the payload. Since Gemini does not have a way to indicate at the beginning the payload size, this will be obtained only if `get_content`is true, and it will be limited by the parameter `maxsize`
`mediatype`: the media type (MIME type) of the resource, such as `text/gemini` or `image/jpeg`
`lang`: the human language of the resource, as standardized in BCP 47
`charset`: actually the encoding of the resource such as UTF-8 or US-ASCII
`tls_version`: the TLS version, for instance, "TLSv1.3"
`no_shutdown`: set to True if the server did not properly close the TLS session. It may mean that the content was truncated. Meaningful only with `get_content=True` and if you asked for the whole file.
The rest is related to certificates:
`issuer`: the CA (Certificate Authority)
`subject`: the name in the certificate (X.509 calls it "subject")
`cert_not_after`: expiration date
`cert_not_before`: inception date
`cert_algo`: algorithm used by the CA
`cert_key_type`: algorithm of the public key
`keystring`: the public key
`cert_key_size`: size of the public key

BCP 47 (RFC on codes for languages)

See `sample-client.py`. (In the source code, the test suite under `tests/` is also a good way to learn about how to use the library.) Agunua is used in the Manisha monitoring tool and in the Lupa crawler.

Manisha

Lupa

Command-line client

`agunua` is a simple Gemini command-line client, a bit like curl. Most parameters of the library `GeminiUri()` constructor can be set via options. Important: the default value is not always the same with the command-line tool. For instance, it defaults to actually retrieving the content.

Details on the command-line client

Download an entire capsule

Another command-line client, `geminitrack`, allows you to retrieve an entire capsule, for instance for backups.

Details on geminitrack

Name

Agunua is a melanesian serpent god. Caduceus would have been better for a Python + Gemini project since there are two snakes on a caduceus but it was already used.

License

GPL. See LICENSE.

Authors

Stéphane Bortzmeyer <stephane+framagit@bortzmeyer.org>.

Reference site

Gitlab site on the Web.

Use the Gitlab issue tracker to report bugs or wishes.

Reference site on Gemini

Other Gemini clients in Python