Working on tree docs.

This commit is contained in:
Simon Forman 2018-05-29 11:38:58 -07:00
parent aafecdc035
commit e67a6f7b6e
5 changed files with 2829 additions and 7 deletions

View File

@ -2979,7 +2979,7 @@
"name": "python", "name": "python",
"nbconvert_exporter": "python", "nbconvert_exporter": "python",
"pygments_lexer": "ipython2", "pygments_lexer": "ipython2",
"version": "2.7.13" "version": "2.7.12"
} }
}, },
"nbformat": 4, "nbformat": 4,

View File

@ -258,7 +258,7 @@
"name": "python", "name": "python",
"nbconvert_exporter": "python", "nbconvert_exporter": "python",
"pygments_lexer": "ipython2", "pygments_lexer": "ipython2",
"version": "2.7.13" "version": "2.7.12"
} }
}, },
"nbformat": 4, "nbformat": 4,

File diff suppressed because it is too large Load Diff

View File

@ -1482,14 +1482,14 @@
"source": [ "source": [
"# BTree-delete\n", "# BTree-delete\n",
"\n", "\n",
"Now let's write a function that can return a tree datastructure with a key, value pair deleted:\n",
"\n", "\n",
" tree key [Er] BTree-delete\n", " tree key BTree-delete\n",
" -------------------------------- key in tree\n", " ---------------------------\n",
" tree\n", " tree\n",
"\n", "\n",
" tree key [Er] BTree-delete\n", "\n",
" ---------------------- ------ key not in tree\n", "If the key is not in tree it just returns the tree unchanged."
" tree key Er"
] ]
}, },
{ {

692
docs/Treestep.ipynb Normal file
View File

@ -0,0 +1,692 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Treating Trees II\n",
"Let's consider a tree structure, similar to one described [\"Why functional programming matters\" by John Hughes](https://www.cs.kent.ac.uk/people/staff/dat/miranda/whyfp90.pdf), that consists of a node value followed by a sequence of zero or more child trees. (The asterisk is meant to indicate the [Kleene star](https://en.wikipedia.org/wiki/Kleene_star).)\n",
"\n",
" tree = [] | [node tree*]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## `treestep`\n",
"In the spirit of `step` we are going to define a combinator `treestep` which expects a tree and three additional items: a base-case value `z`, and two quoted programs `[N]` and `[C]`.\n",
"\n",
" tree z [N] [C] treestep"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The base-case value `z` is of some type `Z`, the `[N]` function is executed per-node and should accept the node value (whatever it is) and return a value of type `A` (which can be `Z`), and the `[C]` function should expect an `A` and a list of `Z` and return a value of type `Z`:\n",
"\n",
" z :: Z\n",
" N :: node -> A\n",
" C :: A [Z*] -> Z"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"If the current tree node is empty then just leave `z` on the stack in lieu:\n",
"\n",
" [] z [N] [C] treestep\n",
" ---------------------------\n",
" z"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Otherwise, evaluate `N` on the node value, `map` the whole function (abbreviated here as `K`) over the child trees recursively, and then combine the result with `C`.\n",
"\n",
" [node tree*] z [N] [C] treestep\n",
" --------------------------------------- w/ K == z [N] [C] treestep\n",
" node N [tree*] [K] map C"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Derive the recursive form.\n",
"We can begin to derive it by finding the `ifte` stage that `genrec` will produce. The predicate and base-case functions are trivial, so we just have to derive `J`.\n",
"\n",
" K == [not] [pop z] [J] ifte"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The behavior of `J` is to accept a (non-empty) tree node and arrive at the desired outcome.\n",
"\n",
" [node tree*] J\n",
" ------------------------------\n",
" node N [tree*] [K] map C"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"So `J` will have some form like:\n",
"\n",
" J == .. [N] .. [K] .. [C] .."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's dive in. First, unquote the node and `dip` `N`.\n",
"\n",
" [node tree*] uncons [N] dip\n",
" node [tree*] [N] dip\n",
" node N [tree*]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next, `map` `K` over the child trees and combine with `C`.\n",
"\n",
" node N [tree*] [K] map C\n",
" node N [tree*] [K] map C\n",
" node N [K.tree*] C"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"So:\n",
"\n",
" J == uncons [N] dip [K] map C"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Plug it in and convert to `genrec`:\n",
"\n",
" K == [not] [pop z] [J ] ifte\n",
" K == [not] [pop z] [uncons [N] dip [K] map C] ifte\n",
" K == [not] [pop z] [uncons [N] dip] [map C] genrec"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Extract the givens to parameterize the program.\n",
"Working backwards:\n",
"\n",
" [not] [pop z] [uncons [N] dip] [map C] genrec\n",
" [not] [z] [pop] swoncat [uncons [N] dip] [map C] genrec\n",
" [not] z unit [pop] swoncat [uncons [N] dip] [map C] genrec\n",
" z [not] swap unit [pop] swoncat [uncons [N] dip] [map C] genrec\n",
" ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
" TS0 == [not] swap unit [pop] swoncat"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
" z TS0 [uncons [N] dip] [map C] genrec\n",
" z [uncons [N] dip] [TS0] dip [map C] genrec\n",
" z [[N] dip] [uncons] swoncat [TS0] dip [map C] genrec\n",
" z [N] [dip] cons [uncons] swoncat [TS0] dip [map C] genrec\n",
" ^^^^^^^^^^^^^^^^^^^^^^^^^^^"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
" TS1 == [dip] cons [uncons] swoncat"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
" z [N] TS1 [TS0] dip [map C] genrec\n",
" z [N] [map C] [TS1 [TS0] dip] dip genrec\n",
" z [N] [C] [map] swoncat [TS1 [TS0] dip] dip genrec\n",
"\n",
"The givens are all to the left so we have our definition."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Define `treestep`"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [],
"source": [
"from notebook_preamble import D, J, V, define, DefinitionWrapper"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [],
"source": [
"DefinitionWrapper.add_definitions('''\n",
"\n",
" _treestep_0 == [not] swap unit [pop] swoncat\n",
" _treestep_1 == [dip] cons [uncons] swoncat\n",
" treegrind == [_treestep_1 [_treestep_0] dip] dip genrec\n",
" treestep == [map] swoncat treegrind\n",
"\n",
"''', D)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Examples\n",
"Consider trees, the nodes of which are integers. We can find the sum of all nodes in a tree with this function:\n",
"\n",
" sumtree == 0 [] [sum +] treestep"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Running this function on an empty tree value gives zero:\n",
"\n",
" [] 0 [N] [C] treestep\n",
" ---------------------------\n",
" 0"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Running it on a non-empty node:\n",
"\n",
" [n tree*] 0 [] [sum +] treestep\n",
" n [tree*] [0 [] [sum +] treestep] map sum +\n",
" n [ ... ] sum +\n",
" n m +\n",
" n+m\n"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"define('sumtree == 0 [] [sum +] treestep')"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"0\n"
]
}
],
"source": [
"J('[] sumtree') # Empty tree."
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"23\n"
]
}
],
"source": [
"J('[23] sumtree') # No child trees."
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"23\n"
]
}
],
"source": [
"J('[23 []] sumtree') # Child tree, empty."
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"32\n"
]
}
],
"source": [
"J('[23 [2 [4]] [3]] sumtree') # Non-empty child trees."
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"49\n"
]
}
],
"source": [
"J('[23 [2 [8] [9]] [3] [4 []]] sumtree') # Etc..."
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"49\n"
]
}
],
"source": [
"J('[23 [2 [8] [9]] [3] [4 []]] 0 [] [cons sum] treestep') # Alternate \"spelling\"."
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[23 [23 [23] [23]] [23] [23 []]]\n"
]
}
],
"source": [
"J('[23 [2 [8] [9]] [3] [4 []]] [] [pop 23] [cons] treestep') # Replace each node."
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[1 [1 [1] [1]] [1] [1 []]]\n"
]
}
],
"source": [
"J('[23 [2 [8] [9]] [3] [4 []]] [] [pop 1] [cons] treestep')"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"6\n"
]
}
],
"source": [
"J('[23 [2 [8] [9]] [3] [4 []]] [] [pop 1] [cons] treestep sumtree')"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {
"scrolled": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"6\n"
]
}
],
"source": [
"J('[23 [2 [8] [9]] [3] [4 []]] 0 [pop 1] [sum +] treestep') # Combine replace and sum into one function."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Redefining our BTree in terms of this form.\n",
"\n",
" BTree = [] | [[key value] left right]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"What kind of functions can we write for this with our `treestep`?\n",
"\n",
"The pattern for processing a non-empty node is:\n",
"\n",
" node N [tree*] [K] map C\n",
"\n",
"Plugging in our BTree structure:\n",
"\n",
" [key value] N [left right] [K] map C"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Traversal\n",
" [key value] uncons pop [left right] [K] map i\n",
" key [value] pop [left right] [K] map i\n",
" key [left right] [K] map i\n",
" key [lkey rkey ] i\n",
" key lkey rkey"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This doesn't quite work:"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"3 23 23\n"
]
}
],
"source": [
"J('[[3 0] [[2 0] [][]] [[9 0] [[5 0] [[4 0] [][]] [[8 0] [[6 0] [] [[7 0] [][]]][]]][]]] 23 [uncons pop] [i] treestep')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Doesn't work because `map` extracts the `first` item of whatever its mapped function produces. We have to return a list, rather than depositing our results directly on the stack.\n",
"\n",
"\n",
" [key value] N [left right] [K] map C\n",
"\n",
" [key value] first [left right] [K] map flatten cons\n",
" key [left right] [K] map flatten cons\n",
" key [[lk] [rk] ] flatten cons\n",
" key [ lk rk ] cons\n",
" [key lk rk ]\n",
"\n",
"So:\n",
"\n",
" [] [flatten cons] [first] treestep"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[3 2 9 5 4 8 6 7]\n"
]
}
],
"source": [
"J('[[3 0] [[2 0] [] []] [[9 0] [[5 0] [[4 0] [] []] [[8 0] [[6 0] [] [[7 0] [] []]] []]] []]] [] [first] [flatten cons] treestep')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"There we go."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### In-order traversal with `treestep`.\n",
"\n",
"From here:\n",
"\n",
" key [[lk] [rk]] C\n",
" key [[lk] [rk]] i\n",
" key [lk] [rk] roll<\n",
" [lk] [rk] key swons concat\n",
" [lk] [key rk] concat\n",
" [lk key rk]\n",
"\n",
"So:\n",
"\n",
" [] [i roll< swons concat] [first] treestep"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[2 3 4 5 6 7 8 9]\n"
]
}
],
"source": [
"J('[[3 0] [[2 0] [] []] [[9 0] [[5 0] [[4 0] [] []] [[8 0] [[6 0] [] [[7 0] [] []]] []]] []]] [] [uncons pop] [i roll< swons concat] treestep')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## with `treegrind`?\n",
"What kind of functions can we write for this with our `treegrind`?\n",
"\n",
"The pattern for processing a non-empty node is:\n",
"\n",
" node N [tree*] [K] C\n",
"\n",
"Plugging in our BTree structure:\n",
"\n",
" [key value] N [left right] [K] C"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"23 'N' [[44] [18]] [[not] [pop 'z'] [uncons ['N'] dip] ['C'] genrec] 'C'\n"
]
}
],
"source": [
"J('[23 [44] [18] ] \"z\" [\"N\"] [\"C\"] treegrind')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[3 0] 'N' [2 0] 'N' 23 23 [9 0] 'N' [5 0] 'N' [4 0] 'N' 23 23 [8 0] 'N' [6 0] 'N' 23 [7 0] 'N' 23 23 23 23\n"
]
}
],
"source": [
"J('[[3 0] [[2 0] [] []] [[9 0] [[5 0] [[4 0] [] []] [[8 0] [[6 0] [] [[7 0] [] []]] []]] []]] 23 [\"N\"] [step] treegrind')"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"0 3 2 23 23 9 5 4 23 23 8 6 23 7 23 23 23 23\n"
]
}
],
"source": [
"J('0 [[3 0] [[2 0] [] []] [[9 0] [[5 0] [[4 0] [] []] [[8 0] [[6 0] [] [[7 0] [] []]] []]] []]] 23 [first ] [step] treegrind')"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"0 [3 [2 [] []] [9 [5 [4 [] []] [8 [6 [] [7 [] []]] []]] []]]\n"
]
}
],
"source": [
"J('0 [[3 0] [[2 0] [] []] [[9 0] [[5 0] [[4 0] [] []] [[8 0] [[6 0] [] [[7 0] [] []]] []]] []]] [] [first ] [map cons] treegrind')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"language": "python",
"name": "python2"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.12"
}
},
"nbformat": 4,
"nbformat_minor": 2
}