Folgezettel tree improvements

guide/2011503.md

@@ -1,12 +1,15 @@
 # Graph view
 
-A zettelkasten is a [directed graph](https://en.wikipedia.org/wiki/Directed_graph), and <2017401> is a subset of this graph established by having zettels branch off to other zettels.
+A zettelkasten is a [directed graph](https://en.wikipedia.org/wiki/Directed_graph). Neuron also has the notion of <2017401>, which is a subset of this graph established by having zettels "branch off" to other zettels.
 
 ## z-index 
 
-The z-index page (at `/z-index.html`; also linked in the footer) displays your Zettelkasten folgezettel graph. Neuron detects if there are any cycles in the graph (use `cf` to resolve cycles). Then, it detects all <2012301> in the graph, and displays the <2017401> for each cluster.
+The z-index page (at `/z-index.html`; also linked in the header) displays the zettelkasten <2017401> for all <2012301> in the zettelkasten graph.
 
 ## Uplinks and Backlinks
 
-Uplinks are a kind of backlinks. Specifically an **uplink tree** of a zettel is the subset of the <2017401> which branch off to the zettel. Uplink tree is displayed above the zettel; other backlinks are displayed below.
+A backlink of a zettel is a zettel that links to it. If that link is a folgezettel link, it is called an "uplink". 
 
+An **uplink tree** of a zettel is the subset of the <2017401> which branch off to the zettel. Uplink tree is displayed above the zettel; other backlinks are displayed below.
+
+Use `?cf` (see <2011504?cf>) to eject something out of a zettel's uplink tree.

guide/2011504.md

@@ -18,13 +18,22 @@ linked zettel.
 
 If your link is merely a reference to another zettel, and you do not wish it to
 be part of the <2017401?cf>, but only the graph connection, use the `cf`[^cf] query
-flag (eg: `<ef3dke98?cf>`). Neuron will link the zettels, but the link would be
-ignored when building the <2017401?cf>. This is especially useful to resolve
-cycles and ensure sensible clusters.
+flag (eg: `<ef3dke98?cf>`). 
+
+```markdown
+This is a zettel file, which refers to another zettel without 
+strongly linking to it:
+
+* <ef3dke98?cf>
+```
+
+Neuron will link the zettels, but the link would be
+ignored when building the <2017401?cf>. This is especially useful to  ensure sensible
+clusters.
 
 ## Advanced linking
 
-* [[2011506]]
+* <2011506>
 
 [^cf]: 
     > The abbreviation cf. (short for the Latin: *confer/conferatur*, both meaning "compare") is used in writing to refer the reader to other material to make a comparison with the topic being discussed. [Wikipedia](https://en.wikipedia.org/wiki/Cf.)

guide/index.md

@@ -13,7 +13,7 @@
 
 ## Cerveau
 
-[Cerveau](https://www.cerveau.app) is the upcoming web and mobile app ([see screenshot](https://lh3.googleusercontent.com/-aBX5ePS0mD6Ey8NQUFhRmX8SHhxr0aHBE9rX4aGPU_i4_NloQsJsSP2Ug0D8oJGrQeyz2H3MYrnyoJC83mGOsoHOkfB_z6kqDJUa7bbO2xHATw9pw-cKvvMkNpMnW1ww7Z-gpCaM4nm7yTFLSa9T7YQYdWOGJVnBy-vPBRX6d_3kJ0JZj9WoOBQtYnOUiLD85tEjF03PtVVw9mSvhpqrTPLicThyAHj3koYEJAXtdWqP_X7s4lSHvVqtod9NfvM8WFNSsITZJmBHbP8bfGBs1aVy2zCM7cU63bsMw5jVAA5wXm6Ef9v9Lf2gqY-SgRimzlqnUX4eAfEnhvmAkGEOiQgyJ8QJSSgm99TW9l_ve7ZQPZmXgnfUHHCjYwIoqc72VXjTEMn-RRhRJDD8txb3mRQU05t-nF7U4dW8XtWuyXFxB-7Vc7lkwofSEWh_3dJiPZwIT221qFCxz8pr8krMEPrzgf_-YSFoGo804PTXGpp0N8_gM4ElTt4KXx3jfyYiGx4OUD9Lm_n-FaLFyWas7ryIKmTDmx78t1PhycCjwoBT3k3BwWLk-RIcSV9KvCKpkhC2p2wf8qC5ekPgaGaj6J2Rh-OrAI5zF9-erRCzwVw6NF2E-xQFg_8VmIp4Lpvl1ygwB4pX9xKlq4qrGyxVtxO2bIcwnrELqx81Yi_Oah8Y2ik0_VVAdEkrzeDOX4=w3381-h2269-no?authuser=0)) for easy editing and publishing of Neuron notes backed by [Git](https://guides.github.com/introduction/git-handbook/) as storage.
+[Cerveau](https://www.cerveau.app) is the upcoming web and mobile app for easy editing and publishing of Neuron notes backed by [Git](https://guides.github.com/introduction/git-handbook/) as storage.
 
 ## External links
 

neuron/neuron.cabal

@@ -1,7 +1,7 @@
 cabal-version: 2.4
 name: neuron
 -- This version must be in sync with what's in Default.dhall
-version: 0.5.6.3
+version: 0.5.7.0
 license: AGPL-3.0-only
 copyright: 2020 Sridhar Ratnakumar
 maintainer: srid@srid.ca

neuron/src/lib/Data/Graph/Labelled.hs

@@ -9,6 +9,7 @@ module Data.Graph.Labelled
     mkGraphFrom,
 
     -- * Querying
+    getGraph,
     findVertex,
     getVertices,
     hasEdge,
@@ -22,6 +23,8 @@ module Data.Graph.Labelled
     clusters,
     dfsForestFrom,
     dfsForestBackwards,
+    bfsForestFrom,
+    bfsForestBackwards,
     obviateRootUnlessForest,
     induceOnEdge,
   )

neuron/src/lib/Data/Graph/Labelled/Algorithm.hs

@@ -16,6 +16,10 @@ import qualified Data.Set as Set
 import Data.Tree (Forest, Tree (..))
 import Relude
 
+{-# INLINE getGraph #-}
+getGraph :: LabelledGraph v e -> LAM.AdjacencyMap e (VertexID v)
+getGraph (LabelledGraph g _) = g
+
 findVertex :: Ord (VertexID v) => VertexID v -> LabelledGraph v e -> Maybe v
 findVertex v lg@(LabelledGraph g _) = do
   guard $ LAM.hasVertex v g
@@ -88,6 +92,14 @@ dfsForestBackwards :: (Monoid e, Vertex v, Ord (VertexID v)) => v -> LabelledGra
 dfsForestBackwards fromV (LabelledGraph g' v') =
   dfsForestFrom [fromV] $ LabelledGraph (LAM.transpose g') v'
 
+bfsForestBackwards :: (Monoid e, Vertex v, Ord (VertexID v)) => v -> LabelledGraph v e -> Forest v
+bfsForestBackwards fromV (LabelledGraph g' v') =
+  bfsForestFrom [fromV] $ LabelledGraph (LAM.transpose g') v'
+
+bfsForestFrom :: (Vertex v, Ord (VertexID v)) => [v] -> LabelledGraph v e -> Forest v
+bfsForestFrom (fmap vertexID -> vs) g =
+  fmap (fmap $ getVertex g) $ Algo.bfsForest vs $ LAM.skeleton $ graph g
+
 --------------------------
 --- More general utilities
 --------------------------

neuron/src/lib/Neuron/Web/ZIndex.hs

@@ -36,23 +36,27 @@ import Relude hiding ((&))
 -- All heavy graph computations are decoupled from rendering, producing this
 -- value, that is in turn used for instant rendering.
 data ZIndex = ZIndex
-  { -- | Topological ordering of the folgezettel graph. Left value indicates
-    --  that it is cyclic.
-    zIndexTopSort :: Either (NonEmpty Zettel) [Zettel],
-    -- | Clusters on the folgezettel graph.
+  { -- | Clusters on the folgezettel graph.
     zIndexClusters :: [Forest (Zettel, [Zettel])],
     -- | All zettel errors
-    zIndexErrors :: Map ZettelID ZettelError
+    zIndexErrors :: Map ZettelID ZettelError,
+    zIndexStats :: Stats
   }
 
+data Stats = Stats
+  { statsZettelCount :: Int,
+    statsZettelConnectionCount :: Int
+  }
+  deriving (Eq, Show)
+
 buildZIndex :: ZettelGraph -> Map ZettelID ZettelError -> ZIndex
 buildZIndex graph errors =
   let clusters = G.categoryClusters graph
       clusters' :: [Forest (Zettel, [Zettel])] =
         flip fmap clusters $ \(zs :: [Tree Zettel]) ->
           G.backlinksMulti Folgezettel zs graph
-      topSort = G.topSort graph
-   in ZIndex topSort (fmap sortCluster clusters') errors
+      stats = Stats (length $ G.getZettels graph) (G.connectionCount graph)
+   in ZIndex (fmap sortCluster clusters') errors stats
   where
     -- TODO: Either optimize or get rid of this (or normalize the sorting somehow)
     sortCluster fs =
@@ -70,14 +74,13 @@ renderZIndex (Theme.semanticColor -> themeColor) ZIndex {..} = do
   elClass "h1" "header" $ text "Zettel Index"
   renderErrors zIndexErrors
   divClass "z-index" $ do
-    case zIndexTopSort of
-      Left (toList -> cyc) -> divClass "ui orange segment" $ do
-        el "h2" $ text "Cycle detected"
-        forM_ cyc $ \zettel ->
-          el "li" $ QueryView.renderZettelLink Nothing def zettel
-      _ -> blank
     el "p" $ do
-      text $ "There " <> countNounBe "cluster" "clusters" (length zIndexClusters) <> " in the Zettelkasten folgezettel graph. "
+      text $
+        "The zettelkasten has "
+          <> countNounBe "zettel" "zettels" (statsZettelCount zIndexStats)
+          <> " and "
+          <> countNounBe "link" "links" (statsZettelConnectionCount zIndexStats)
+      text $ ". It has " <> countNounBe "cluster" "clusters" (length zIndexClusters) <> " in its folgezettel graph. "
       text "Each cluster's "
       elAttr "a" ("href" =: "https://neuron.zettel.page/2017401.html") $ text "folgezettel heterarchy"
       text " is rendered as a forest."
@@ -86,8 +89,8 @@ renderZIndex (Theme.semanticColor -> themeColor) ZIndex {..} = do
         el "ul" $ renderForest forest
   where
     countNounBe noun nounPlural = \case
-      1 -> "is 1 " <> noun
-      n -> "are " <> show n <> " " <> nounPlural
+      1 -> "1 " <> noun
+      n -> show n <> " " <> nounPlural
 
 renderErrors :: DomBuilder t m => Map ZettelID ZettelError -> NeuronWebT t m ()
 renderErrors errors = do

neuron/src/lib/Neuron/Zettelkasten/Graph.hs

@@ -21,9 +21,11 @@ module Neuron.Zettelkasten.Graph
     backlinksMulti,
     clusters,
     categoryClusters,
+    connectionCount,
   )
 where
 
+import qualified Algebra.Graph.Labelled.AdjacencyMap as LAM
 import Data.Default
 import Data.Foldable (maximum)
 import qualified Data.Graph.Labelled as G
@@ -37,13 +39,13 @@ import Relude
 frontlinkForest :: Connection -> Zettel -> ZettelGraph -> Forest Zettel
 frontlinkForest conn z =
   G.obviateRootUnlessForest z
-    . G.dfsForestFrom [z]
+    . G.bfsForestFrom [z]
     . G.induceOnEdge (== Just conn)
 
 backlinkForest :: Connection -> Zettel -> ZettelGraph -> Forest Zettel
 backlinkForest conn z =
   G.obviateRootUnlessForest z
-    . G.dfsForestBackwards z
+    . G.bfsForestBackwards z
     . G.induceOnEdge (== Just conn)
 
 backlinks :: Connection -> Zettel -> ZettelGraph -> [Zettel]
@@ -67,7 +69,7 @@ backlinksMulti conn zs g =
 categoryClusters :: ZettelGraph -> [Forest Zettel]
 categoryClusters (categoryGraph -> g) =
   let cs :: [[Zettel]] = sortMothers $ clusters g
-   in flip fmap cs $ \zs -> G.dfsForestFrom zs g
+   in flip fmap cs $ \zs -> G.bfsForestFrom zs g
   where
     -- Sort clusters with newer mother zettels appearing first.
     sortMothers :: [NonEmpty Zettel] -> [[Zettel]]
@@ -96,3 +98,6 @@ getZettel = G.findVertex
 -- | If no connection exists, this returns Nothing.
 getConnection :: Zettel -> Zettel -> ZettelGraph -> Maybe Connection
 getConnection z1 z2 g = fmap (fromMaybe def) $ G.edgeLabel g z1 z2
+
+connectionCount :: ZettelGraph -> Int
+connectionCount = LAM.edgeCount . G.getGraph

neuron/test/Neuron/VersionSpec.hs

@@ -28,10 +28,10 @@ spec = do
       "0.4" `isLesserOrEqual` olderThan
     it "full versions" $ do
       "0.6.1.2" `isGreater` olderThan
-      "0.5.8" `isGreater` olderThan
-      "0.5.6.8" `isGreater` olderThan
-      "0.5.6.0" `isLesserOrEqual` olderThan -- This is current version
+      "0.5.9" `isGreater` olderThan
+      "0.5.7.8" `isGreater` olderThan
+      "0.5.7.0" `isLesserOrEqual` olderThan -- This is current version
       "0.3.1.0" `isLesserOrEqual` olderThan
     it "within same major version" $ do
-      "0.5.6.8" `isGreater` olderThan
-      "0.5.6.0" `isLesserOrEqual` olderThan -- This is current version
+      "0.5.7.8" `isGreater` olderThan
+      "0.5.7.0" `isLesserOrEqual` olderThan -- This is current version