{ "slug": "irc-log-parser", "name": "gizza-ai/irc-log-parser", "version": "0.1.0", "title": "IRC Log Parser — gizza.ai", "description": "Parse raw IRC logs from WeeChat, irssi, mIRC, ZNC or HexChat into a readable timeline, or export them as JSON, NDJSON, CSV or a Markdown table.", "tags": [ "irc", "irc log", "weechat", "irssi", "hexchat", "mirc", "znc", "log parser", "chat log", "timeline" ], "category": "developer", "urls": { "page": "https://gizza.ai/tools/irc-log-parser/", "markdown": "https://gizza.ai/tools/irc-log-parser/index.md", "descriptor": "https://gizza.ai/tools/irc-log-parser/tool.json", "deep_link_example": "https://gizza.ai/tools/irc-log-parser/?log=---%20Log%20opened%20Fri%20Jan%2005%2020%3A00%3A00%202024%0A21%3A07%20%3Calice%3E%20shipping%20the%20parser%20tonight%0A21%3A07%20-%21-%20bob%20%5B~bob%40example.net%5D%20has%20joined%20%23gizza%0A21%3A11%20-%21-%20alice%20%5B~a%40example.net%5D%20has%20quit%20%5BPing%20timeout%3A%20240%20seconds%5D&format=auto&output=timeline&date=2024-01-05%20%28for%20logs%20with%20a%20time%20only%29&time_format=iso&include=all&nicks=alice%2C%20bob%2A%20%28leave%20empty%20for%20everyone%29&channel=%23gizza%20%28for%20logs%20that%20never%20name%20it%29&strip_formatting=true&include_raw=true" }, "cli": "gizza tool irc-log-parser \"--- Log opened Fri Jan 05 20:00:00 2024\n21:07 shipping the parser tonight\n21:07 -!- bob [~bob@example.net] has joined #gizza\n21:11 -!- alice [~a@example.net] has quit [Ping timeout: 240 seconds]\"", "tool": { "description": "Turn a raw IRC client log into a structured, readable timeline or a machine-readable export. It reads the timestamp grammars of the common clients — WeeChat (date TAB nick TAB text), irssi (bare '21:07'), mIRC/ZNC/EnergyMech ('[21:07:33]'), HexChat/XChat ('Jan 05 21:07:33'), plain ISO datetimes, and logs with no timestamps at all — auto-detecting by default, and understands each client's event wording regardless of which grammar was used: '-!- alice [~a@host] has joined #chan', '*** Joins: alice (~a@host)', '* Parts: alice (~a@host) (Leaving)', 'alice was kicked from #chan by bob', 'alice is now known as bobby', 'mode/#chan [+o bob] by alice' and the topic phrasings. Every line becomes a typed record — message, action, notice, join, part, quit, kick, nick, mode, topic, meta or unknown — with the same eight fields: line, time, type, nick, host, channel, arg (the new nick, the kicker, or the mode string) and text. Render them as a readable timeline, a JSON array, NDJSON, CSV or a Markdown table; normalize timestamps to ISO, 24-hour, 12-hour, the original text or nothing; supply a base date for time-only logs (irssi day-change markers roll it forward automatically); attach a channel name to lines that lack one; keep only messages or only events; filter to particular nicks with a trailing * for prefix matching; strip mIRC colour and formatting codes; optionally keep the original raw line; and cap the number of records. Fully local and deterministic — no AI model, no network, nothing uploaded.", "parameters": { "additionalProperties": false, "properties": { "channel": { "default": "", "description": "Channel name to record on lines that do not name one themselves, e.g. '#gizza'. Must start with # & or !. Empty by default, which leaves the channel column blank on those lines.", "type": "string" }, "date": { "default": "", "description": "Base calendar date as YYYY-MM-DD for logs that only record a time of day, e.g. '2024-01-05'. Irssi '--- Log opened' and '--- Day changed' markers in the log override it from that point on. Empty by default, which leaves such records with a time only.", "type": "string" }, "format": { "default": "auto", "description": "Timestamp grammar of the log: auto (detect it from the first 200 lines), weechat (date TAB nick TAB text), irssi (bare '21:07' or '21:07:33'), bracket ('[21:07:33]' — mIRC, ZNC, EnergyMech, Textual), hexchat ('Jan 05 21:07:33'), iso ('2024-01-05 21:07:33'), or plain (no timestamps). Event wording from every client is understood in all modes. Default auto.", "enum": [ "auto", "weechat", "irssi", "bracket", "hexchat", "iso", "plain" ], "type": "string" }, "include": { "default": "all", "description": "Which lines to keep: all, messages (message, action and notice lines only), or events (join, part, quit, kick, nick change, mode and topic only). Default all.", "enum": [ "all", "messages", "events" ], "type": "string" }, "include_raw": { "default": false, "description": "Add the original untouched log line as a 'raw' field (json, ndjson) or column (csv). Off by default to keep the output compact.", "type": "boolean" }, "limit": { "default": 0, "description": "Maximum number of records to return, applied after the include and nicks filters. 0 (the default) means no limit; the ceiling is 200000.", "maximum": 200000, "minimum": 0, "type": "integer" }, "log": { "description": "The raw IRC log text. Paste the lines straight out of the client, e.g. '21:07 hey' (irssi), '2024-01-05 21:07:33alicehey' (WeeChat), '[21:07:33] *** Joins: alice (~a@host)' (ZNC/mIRC) or 'Jan 05 21:07:33 hey' (HexChat). Maximum 5000000 bytes.", "type": "string" }, "nicks": { "default": "", "description": "Comma-separated nicks to keep, matched case-insensitively — e.g. 'alice, bob*'. A trailing * matches by prefix, so 'bob*' keeps bob and bobby. Lines with no nick (server notices, log markers) are dropped when this is set. Empty by default, which keeps everyone.", "type": "string" }, "output": { "default": "timeline", "description": "Result shape: timeline (readable one line per event), json (array of records), ndjson (one compact record per line for jq or streaming), csv (spreadsheet columns line,time,type,nick,host,channel,arg,text), or markdown (a table). Default timeline.", "enum": [ "timeline", "json", "ndjson", "csv", "markdown" ], "type": "string" }, "strip_formatting": { "default": true, "description": "When true (default), remove mIRC formatting codes (bold, italic, underline, reverse, reset and ^C colour pairs) and ANSI escape sequences from the text. Turn it off to keep the control characters byte-for-byte.", "type": "boolean" }, "time_format": { "default": "iso", "description": "How each timestamp is written out: iso ('2024-01-05T21:07:33', falling back to '21:07:33' when no date is known), 24h ('21:07:33'), 12h ('9:07:33 PM'), original (exactly as it appeared in the log), or none (drop timestamps). Default iso.", "enum": [ "iso", "24h", "12h", "original", "none" ], "type": "string" } }, "required": [ "log" ], "type": "object" } } }