Initial commit

author: self <george@gabbott.dev> 2023-10-02 20:35:53 +0100
committer: George Abbott <george@gabbott.dev> 2023-10-02 20:36:48 +0100
commit: 8e4092347975db7fe90b0895f605bdd79de47410 (patch)
tree: e7aa55355f4f72b9f7972077539e9e05b6e50702
3 files changed, 221 insertions, 0 deletions
diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..086d612
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,16 @@
+.dub
+docs.json
+__dummy.html
+docs/
+/reassess
+reassess.so
+reassess.dylib
+reassess.dll
+reassess.a
+reassess.lib
+reassess-test-*
+*.exe
+*.pdb
+*.o
+*.obj
+*.lst
diff --git a/dub.json b/dub.json
new file mode 100644
index 0000000..e2848da
--- /dev/null
+++ b/dub.json
@@ -0,0 +1,14 @@
+{
+	"name": "reassess",
+	"copyright": "Copyright © 2024, George",
+	"license": "GPL-3.0",
+	"description": "Generates an RSS feed given a directory of entries.",
+
+	"authors": [
+		"George Abbott"
+	],
+	"homepage": "https://gabbott.dev/project/reassess",
+	"dependencies": {
+		"dxml": "~>0.4.4"
+	}
+}
diff --git a/source/app.d b/source/app.d
new file mode 100644
index 0000000..3c913c3
--- /dev/null
+++ b/source/app.d
@@ -0,0 +1,191 @@
+/* reassess: generates an RSS feed given a directory.
+   To run, simply pass in the directory containing all entries, and the
+   intended target RSS location. For instance:
+      reassess $WEBSITE_URL/blog/entries $WEBSITE_URL/blog/rss.xml
+   Then, you are set! Remember to re-run this program whenever an update is
+   made to the website.
+
+   The location of the metadata can be derived either from the file itself, or
+   from an rssmeta.xml file, which should contain all the metadata for the
+   relevant files.
+   If not, please provide the metadata within the file. 
+
+   reassess requires the following bits of metadata to be provided:
+   	 TAG       derived from?
+	 ---------|-------------
+   - <title>       rssmeta:title OR <!-- title: --> OR <title> OR first <h1> tag
+   - <link>        rssmeta:link OR assessrc:global_link OR <!-- link: ... -->
+   - <pubDate>     rssmeta:pubDate OR <!-- pubDate: ... --> OR <!-- written: ... -->
+   - <guid>        rssmeta:guid OR <!-- guid: ... --> OR the same as <link>.
+   - <description> Everything within <body>.
+	
+   It is best to provide an rssmeta.xml or rssmeta.json file, but failing this
+   all information can be derived from the body of the document if correctly
+   written.
+
+   reassess will fail if: 
+   - There is both an rssmeta.xml and an rssmeta.json file in $1.
+   - The rssmeta file is corrupted or incorrectly formatted.
+   - There is insufficient information to populate a metadata field.
+   etc.
+   In all these cases, the program will output to stderr and return a non-zero
+   return.
+
+   Optional Flags:
+   -m	Gives the location of the rssmeta.xml or rssmeta.json if it is not
+   		present in the directory to be converted.
+		`reassess $WEBSITE_URL/blog/entries $WEBSITE_URL/blog/rss.xml -m $METAFILE`
+   -i	Errors which to ignore.
+   		By default, nothing is ignored.
+   -o	Order: in-order; reverse; title-alphabetical; more TBA. 
+    	Default is reverse, which gives most recent entry first.
+
+	Format of rssmeta.xml:
+	<entry>
+		<filename>...</filename>
+		... the metadata entries, e.g. <title>, <link>, etc.
+	</entry>
+	The <filename> attribute is used to find the file in particular. An error
+	is thrown if:
+	- There exists a file in the directory which is not present in the rssmeta
+	  file (You can pass in `-i file-wo-entry` to ignore this check, and skip
+	  said file when outputting RSS.)
+	- There exists a file in the rssmeta file which is not in the directory.
+	  (You can pass in `-i entry-wo-file` to ignore this. 
+
+
+   */
+
+void usage()
+{
+	writeln(
+			"reassess: create an RSS feed out of HTML documents.\n"
+			"Usage: \n"
+			"\treassess DIRECTORY TARGET (FLAGS)\n"
+			"DIRECTORY: The directory where the HTML files are sourced.\n"
+			"TARGET:    The desired target location of the RSS file.\n"
+			"Flags:\n"
+			"\t-h, --help      Display this help message.\n"
+			"\t-m, --metafile  The path of the metafile if not in DIRECTORY.\n"
+			"\t-r, --recursive Whether to recurse DIRECTORY or not.\n"
+			"\t-f, --follow    Whether to follow symlinks in DIRECTORY or not.\n"
+			"\t-o, --order     The order to output the RSS entries.\n"
+			"\t                Options: reverse; forward; alphabetic-by-title.\n"
+			"\t                Default to `reverse` if unspecified.\n"
+			"\t-c, --config    Options to configure functionality. Details TODO\n"
+		   );
+	exit(0);
+}
+
+
+
+struct Entry
+{
+	string title;
+	string link;
+	string pubDate;
+	string guid;
+	string description; // The actual entry itself.
+
+private:
+	int pubDateAsUnixTimestamp; // So that we can sort easily.
+}
+
+Entry make_entry(DirEntry dirEntry)
+{
+	// From the DirEntry, we can get all file info, and also read the full file
+	// since we know its path. We just need to remember, when we read the
+	// entry, to only read between the <body> and </body> tags, whilst
+	// searching the full file for content, and replace all bad characters with
+	// escaped ones.
+	// Or, if <body> does not work, we could try between some comments. 
+	// I say, <!-- bct:beg --> and <!-- bct:end --> would meet those criteria.
+}
+
+
+
+
+import std.stdio;
+import std.algorithm : canFind;
+import std.file : dirEntries, DirEntry;
+
+// Flags which are passed to -i (--ignore).
+immutable FILE_WO_ENTRY = "file-wo-entry";
+immutable ENTRY_WO_FILE = "entry-wo-file";
+immutable DASH_I_PARAMS = [ FILE_WO_ENTRY, ENTRY_WO_FILE ];
+
+int main(string[] args)
+{
+	/* Flags */
+	string metafile;
+	bool entries_set = false;
+	string entries;
+	string target;
+	string[] ignore;
+	bool recursive = false;
+	bool follow_symlinks = false;
+
+	for (int i = 0; i < args.length; ++i)
+	{
+
+		if (args[i] == "-m" || args[i] == "--metafile")
+			metafile = args[++i];
+		else if (args[i] == "-h" || args[i] == "--help")
+			usage();
+		else if (args[i] == "-r" || args[i] == "--recursive")
+			recursive = true;
+		else if (args[i] == "-f" || args[i] == "--follow")
+			follow_symlinks = true;
+		else if (args[i] == "-o" || args[i] == "--order")
+		else if (args[i] == "-i" || args[i] == "--ignore")
+		{
+			auto param = args[++i];
+			if (DASH_I_PARAMS.canFind(param))
+				ignore ~= param;
+			else 
+			{
+				writeln("Bad param for -i:", param);
+				exit(-1);
+			}
+
+		}
+		else 
+		{
+			if (!entries_set)
+			{
+				entries_set = true;
+				entries = metafile[i];
+			}
+			else
+				target = metafile[i];
+		}
+	}
+
+	// Grab list of files in directory.
+	DirEntry[] objects = []; 
+	foreach (DirEntry file; dirEntries(entries, ".html", 
+				recursive? SpanMode.depth : SpanMode.shallow,
+				follow_symlinks)
+			.filter!(file => f.isFile())
+			) 
+	{
+		objects ~= file;
+	}
+
+
+	// Convert the file, given all info for the metadata, into an Entry. 
+	Entry[] entries;
+	foreach (DirEntry dirEntry; objects)
+	{
+		entries ~= make_entry(dirEntry);
+	}
+
+
+	// Iterate over all entries in the order as described by -o flag, and 
+	// create the entry for them. This gets spat back out as a big string.
+	auto entries = create_entries(...);
+
+	// Now, write all the fluff around the entries, and jobs-almost-agoodun.
+
+	// Finally, write back the full RSS XML into the file specified by $2.
+}
author	self <george@gabbott.dev>	2023-10-02 20:35:53 +0100
committer	George Abbott <george@gabbott.dev>	2023-10-02 20:36:48 +0100
commit	8e4092347975db7fe90b0895f605bdd79de47410 (patch)
tree	e7aa55355f4f72b9f7972077539e9e05b6e50702