Move pymonkey stuff to monkey/

Importing pymonkey into webmonkey, let's see how this works.

9 files changed, 1096 insertions, 0 deletions

diff --git a/monkey/action.py b/monkey/action.py
new file mode 100755
index 0000000..057ed0e
--- /dev/null
+++ b/monkey/action.py
@@ -0,0 +1,284 @@
+#!/usr/bin/python3
+
+class Action:
+    # type ∈ ['remove', 'insert', 'solve', 'solve_all']
+    # time: absolute elapsed time since the attempt started, in ms
+    # offset: position of the first inserted/removed character
+    # text: inserted/removed text or query
+    # total, passed: number of test cases
+    def __init__(self, type, time, offset=0, text='', total=0, passed=0):
+        self.type = type
+        self.time = time
+        if type == 'insert' or type == 'remove':
+            self.offset = offset
+            self.length = len(text)
+            self.text = text
+        elif type == 'solve' or type == 'solve_all':
+            self.query = text
+        elif type == 'test':
+            self.total = total
+            self.passed = passed
+
+    def __str__(self):
+        s = 't = ' + str(self.time/1000.0) + '  ' + self.type
+        if self.type in ('insert', 'remove'):
+            s += ' "' + self.text.replace('\n', '\\n').replace('\t', '\\t') + '" at ' + str(self.offset)
+        elif self.type in ('solve', 'solve_all'):
+            s += ' "' + self.query + '"'
+        elif self.type == 'test':
+            s += ' {0} / {1}'.format(self.passed, self.total)
+        return s
+
+    # apply this action to text
+    def apply(self, text):
+        if self.type == 'insert':
+            return text[:self.offset] + self.text + text[self.offset:]
+        elif self.type == 'remove':
+            return text[:self.offset] + text[self.offset+self.length:]
+        else:
+            return text
+
+    # reverse the application of this action
+    def unapply(self, text):
+        if self.type == 'insert':
+            return text[:self.offset] + text[self.offset+self.length:]
+        elif self.type == 'remove':
+            return text[:self.offset] + self.text + text[self.offset:]
+        else:
+            return text
+
+# parse log from database into a list of actions, cleaning up some fluff.
+# ignore non-text actions (queries and tests)
+def parse(data):
+    if data == None:
+        return [], []
+
+    actions = []
+    incorrect = set()
+
+    time = 0
+    code = ''
+    i = 0
+    while i < len(data):
+        # parse one action
+        type = data[i]
+        i += 1
+        dt = int(((data[i] << 8) + (data[i+1])) * 100.0)
+        time += dt
+        i += 2
+        if type == 1:   # insert
+            offset = (data[i] << 8) + data[i+1]
+            i += 2
+            length = (data[i] << 8) + data[i+1]
+            i += 2
+            text = data[i:i+length].decode()
+            i += length
+            action = Action('insert', time, offset=offset, text=text)
+        elif type == 2: # remove
+            offset = (data[i] << 8) + data[i+1]
+            i += 2
+            length = (data[i] << 8) + data[i+1]
+            i += 2
+            text = code[offset:offset+length]
+            action = Action('remove', time, offset=offset, text=text)
+        elif type == 3 or type == 4:    # solve / solve all
+            length = (data[i] << 8) + data[i+1]
+            i += 2
+            query = data[i:i+length].decode()
+            i += length
+            act_type = 'solve' + ('_all' if type == 4 else '')
+            action = Action(act_type, time, text=query)
+        elif type == 8: # test
+            total = data[i]
+            i += 1
+            passed = data[i]
+            i += 1
+            action = Action('test', time, total=total, passed=passed)
+        else:
+            # unsupported action type
+            continue
+
+        # skip normalization if this is the first action
+        if actions == []:
+            actions.append(action)
+            code = action.apply(code)
+            continue
+
+        # add to list of actions; modify previously added action if necessary
+        prev = actions[-1]
+
+        # remove superfluous REMOVE action when newline is inserted (due to editor auto-indent)
+        if prev.type == 'remove' and action.type == 'insert' and \
+           action.time == prev.time and \
+           action.offset == prev.offset and action.length > prev.length and \
+           action.text[action.length-prev.length:] == prev.text:
+            # discard last REMOVE action
+            code = prev.unapply(code)
+            actions.pop()
+
+            # replace current action with something better
+            length = action.length - prev.length
+            new = Action('insert', prev.time, offset = prev.offset, text = action.text[:length])
+            actions.append(new)
+            code = new.apply(code)
+
+        # remove superfluous INSERT action when newline is removed (due to editor auto-indent)
+        elif prev.type == 'remove' and action.type == 'insert' and \
+             action.time == prev.time and \
+             action.offset == prev.offset and action.length < prev.length and \
+             prev.text[prev.length-action.length:] == action.text:
+            # discard last INSERT action
+            code = prev.unapply(code)
+            actions.pop()
+
+            # replace current action with something better
+            length = prev.length - action.length
+            new = Action('remove', prev.time, offset = prev.offset, text = prev.text[:length])
+            actions.append(new)
+            code = new.apply(code)
+
+        # discard INSERT/REMOVE pairs (typos)
+        elif prev.type == 'insert' and action.type == 'remove' and \
+             action.time - prev.time < 10000 and \
+             action.offset == prev.offset and action.text == prev.text:
+            # discard last and current actions
+            code = prev.unapply(code)
+            actions.pop()
+
+        # discard REMOVE/INSERT pairs (deleted char then typed back)
+        elif prev.type == 'remove' and action.type == 'insert' and \
+             action.offset == prev.offset and action.text == prev.text:
+            # discard last and current actions
+            code = prev.unapply(code)
+            actions.pop()
+
+        # otherwise, simply append the current action
+        else:
+            actions.append(action)
+            code = action.apply(code)
+
+    return actions
+
+# expand any multi-char actions (does not do anything for the 99%)
+def expand(actions):
+    i = 0
+    while i < len(actions):
+        if actions[i].type == 'insert' and len(actions[i].text) > 1:
+            a = actions.pop(i)
+            for offset in range(len(a.text)):
+                actions.insert(i+offset, Action('insert', a.time, a.offset+offset, a.text[offset]))
+            i += len(a.text)
+        elif actions[i].type == 'remove' and len(actions[i].text) > 1:
+            a = actions.pop(i)
+            for offset in range(len(a.text)):
+                actions.insert(i, Action('remove', a.time, a.offset+offset, a.text[offset]))
+            i += len(a.text)
+        else:
+            i += 1
+
+# each action in parse() result corresponds to single insertion/deletion.
+# this function merges related adjacent actions
+def compress(actions):
+    # first make each edit change exactly one character, for easier handling
+    expand(actions)
+
+    i = 0
+    while i < len(actions)-1:
+        a = actions[i]
+        b = actions[i+1]
+
+        # merge adjacent INSERT actions
+        # +a +b  →  +ab
+        if a.type == 'insert' and b.type == 'insert' and \
+            b.offset == a.offset + a.length: #and b.time - a.time < 10000:
+            a.text += b.text
+            a.length += b.length
+            del actions[i+1]
+
+        # merge adjacent REMOVE actions (two cases: backspace & delete)
+        # -b -a  →  -ab
+        elif a.type == 'remove' and b.type == 'remove' and \
+            b.offset == a.offset - b.length: #and b.time - a.time < 10000:
+            a.text = b.text + a.text
+            a.offset = b.offset
+            a.length += b.length
+            del actions[i+1]
+        # -a -b  →  -ab
+        elif a.type == 'remove' and b.type == 'remove' and \
+             b.offset == a.offset and b.time - a.time < 10000:
+            a.text += b.text
+            a.length += b.length
+            del actions[i+1]
+
+        # merge adjacent INSERT/REMOVE actions
+        # +ab -b  →  +a
+        elif a.type == 'insert' and b.type == 'remove' and \
+             b.offset >= a.offset and b.offset < a.offset + a.length and \
+             b.length == a.offset + a.length - b.offset and b.time - a.time < 10000:
+            del_start = b.offset - a.offset
+            del_end = del_start + b.length
+            a.text = a.text[:del_start] + a.text[del_end:]
+            a.length -= b.length
+            del actions[i+1]
+
+        else:
+            i += 1
+
+# some sample code
+if __name__ == '__main__':
+    import sys
+    if len(sys.argv) < 2:
+        print('usage: ' + sys.argv[0] + ' <database>')
+        sys.exit(1)
+
+    import sqlite3
+    conn = sqlite3.connect(sys.argv[1])
+    conn.text_factory = bytes
+    c = conn.cursor()
+
+    # print all problem ids
+    print('problems:')
+    c.execute('select * from problems')
+    for problem in c.fetchall():
+        # problem = (id, name, description, details, solution, library)
+        # name:     predicate name + arity (e.g. conc/2)
+        # desc:     one-line problem description
+        # details:  detailed problem description
+        # solution: official solution
+        # library:  fact database for testing (e.g. for parent, brother, … relations)
+        print(' ' + str(problem[0]) + '\t' + str(problem[1], encoding='utf-8'))
+    print()
+
+    pid = input('enter problem id: ')
+    c.execute('select id from attempts where problem=?', (pid,))
+    attempts = list(c.fetchall())
+
+    # print all attempt ids for the selected problem
+    print('attempts for problem ' + str(pid) + ':')
+    print(', '.join([str(attempt[0]) for attempt in attempts]))
+    print()
+
+    aid = input('enter attempt id: ')
+    c.execute('select * from attempts where id=?', (aid,))
+    attempt = c.fetchone()
+    # attempt = (id, problem_id, user_id, log, content, done, session)
+    # log:     action sequence log
+    # content: final version for this attempt
+    # done:    did any version of the program pass all tests?
+    # session: irrelevant
+    try:
+        actions = parse(attempt[3])
+        print('read ' + str(len(actions)) + ' actions from log')
+        compress(actions)
+        print('after compression: ' + str(len(actions)) + ' actions')
+        print()
+
+        print('code versions for this attempt:')
+        code = ''
+        for action in actions:
+            code = action.apply(code)
+            print(action)
+            print(code.strip())
+            print()
+    except Exception as ex:
+        sys.stderr.write('Error parsing action log: ' + str(ex))
diff --git a/monkey/db.py b/monkey/db.py
new file mode 100644
index 0000000..0634098
--- /dev/null
+++ b/monkey/db.py
@@ -0,0 +1,44 @@
+#!/usr/bin/python3
+
+import sqlite3
+
+db = sqlite3.connect('misc/solutions.db')
+db.row_factory = sqlite3.Row
+db.text_factory = bytes
+cursor = db.cursor()
+
+def b_to_utf8(bytestring):
+    return str(bytestring, encoding='utf-8')
+
+def get_problem_ids():
+    cursor.execute('SELECT id FROM problems ORDER BY id ASC')
+    return [row['id'] for row in cursor]
+
+def get_problem(pid):
+    cursor.execute('SELECT name, solution, library FROM problems WHERE id=?', (pid,))
+    row = cursor.fetchone()
+    name = b_to_utf8(row['name'])
+    solution = b_to_utf8(row['solution']).replace('\r', '')
+    lib_id = row['library'] if row['library'] else None
+    return name, solution, lib_id
+
+def get_depends(pid):
+    cursor.execute('SELECT dependency FROM depends WHERE problem=?', (pid,))
+    return [r['dependency'] for r in cursor.fetchall()]
+
+def get_library(lid):
+    cursor.execute('SELECT facts FROM libraries WHERE id=?', (lid,))
+    row = cursor.fetchone()
+    return b_to_utf8(row['facts']).replace('\r', '') if row else None
+
+def get_tests(pid):
+    cursor.execute('SELECT query FROM tests WHERE problem=?', (pid,))
+    return [b_to_utf8(row['query']) for row in cursor]
+
+def get_traces(pid):
+    cursor.execute('SELECT * FROM attempts WHERE problem=? AND done=1 ORDER BY id ASC', (pid,))
+    return {(pid, attempt['user']): attempt['log'] for attempt in cursor}
+
+def get_solved(uid):
+    cursor.execute('SELECT problem FROM attempts WHERE user=? AND done=1 ORDER BY problem ASC', (uid,))
+    return [row['problem'] for row in cursor.fetchall()]
diff --git a/monkey/edits.py b/monkey/edits.py
new file mode 100644
index 0000000..fef591a
--- /dev/null
+++ b/monkey/edits.py
@@ -0,0 +1,307 @@
+#!/usr/bin/python3
+
+import collections
+
+from action import expand, parse
+from graph import Node
+from prolog.util import rename_vars, stringify, tokenize
+from util import get_line
+
+# A line edit is a contiguous sequences of actions within a single line. This
+# function takes a sequence of actions and builds a directed acyclic graph
+# where each edit represents one line edit and each node represents a version
+# of some line. The function returns a list of nodes (first element is the
+# root), and sets of submissions (program versions tested by the user) and
+# queries in this attempt.
+def edit_graph(actions, debug=False):
+    # Return values.
+    nodes = [Node([0, 0, ()])]  # Node data: rank (Y), line no. (X), and tokens.
+    submissions = set()         # Program versions at 'test' actions.
+    queries = set()             # Queries run by the student.
+
+    # State variables.
+    leaves = {0: nodes[0]}  # Current leaf node for each line.
+    rank = 1                # Rank (order / y-position) for the next node.
+    code_next = ''          # Program code after applying the current action.
+
+    # Ensure there is a separate action for each inserted/removed character.
+    expand(actions)
+    for action_id, action in enumerate(actions):
+        code = code_next
+        code_next = action.apply(code)
+
+        if action.type == 'test':
+            submissions.add(code)
+
+        elif action.type == 'solve' or action.type == 'solve_all':
+            queries.add(action.query)
+
+        elif action.type == 'insert' or action.type == 'remove':
+            # Number of the changed line.
+            line = code[:action.offset].count('\n')
+            # Current leaf node for this line.
+            parent = leaves[line]
+            # Tokens in this line after applying [action].
+            tokens_next = tuple(tokenize(get_line(code_next, line)))
+
+            # If a new node is inserted, clone each leaf into the next rank.
+            # This makes it easier to print the graph for graphviz; when
+            # analyzing the graph, duplicate nodes without siblings should be
+            # ignored.
+            new_leaves = {}
+
+            if action.text == '\n':
+                if action.type == 'insert':
+                    tokens_next_right = tuple(tokenize(get_line(code_next, line+1)))
+
+                    child_left = Node([rank, line, tokens_next])
+                    parent.add_out(child_left)
+
+                    child_right = Node([rank, line+1, tokens_next_right])
+                    parent.add_out(child_right)
+
+                    # Create new leaf nodes.
+                    for i, leaf in leaves.items():
+                        if i < line:
+                            new_leaves[i] = Node([rank, i, leaf.data[2]])
+                            leaf.add_out(new_leaves[i])
+                        elif i > line:
+                            new_leaves[i+1] = Node([rank, i+1, leaf.data[2]])
+                            leaf.add_out(new_leaves[i+1])
+                    new_leaves[line] = child_left
+                    new_leaves[line+1] = child_right
+
+                elif action.type == 'remove':
+                    parent_right = leaves[line+1]
+
+                    child = Node([rank, line, tokens_next])
+                    parent_right.add_out(child)
+                    parent.add_out(child)
+
+                    # Create new leaf nodes.
+                    for i, leaf in leaves.items():
+                        if i < line:
+                            new_leaves[i] = Node([rank, i, leaf.data[2]])
+                            leaf.add_out(new_leaves[i])
+                        elif i > line+1:
+                            new_leaves[i-1] = Node([rank, i-1, leaf.data[2]])
+                            leaf.add_out(new_leaves[i-1])
+                    new_leaves[line] = child
+            else:
+                # Skip the node if the next action is insert/remove (except \n)
+                # on the same line.
+                if action_id < len(actions)-1:
+                    action_next = actions[action_id+1]
+                    if action_next.type in ('insert', 'remove'):
+                        line_next = code_next[:action_next.offset].count('\n')
+                        if action_next.text != '\n' and line == line_next:
+                            continue
+
+                # Skip the node if it is the same as the parent.
+                if tokens_next == parent.data[2]:
+                    continue
+
+                child = Node([rank, line, tokens_next])
+                parent.add_out(child)
+
+                # Create new leaf nodes.
+                for i, leaf in leaves.items():
+                    if i != line:
+                        new_leaves[i] = Node([rank, i, leaf.data[2]])
+                        leaf.add_out(new_leaves[i])
+                new_leaves[line] = child
+
+            leaves = new_leaves
+            nodes += leaves.values()
+            rank += 1
+
+    return nodes, submissions, queries
+
+# Return all interesting edit paths in the edit graph rooted at [root].
+def get_paths(root, path=tuple(), done=None):
+    if done is None:
+        done = set()
+
+    cur_path = list(path)
+    if len(path) == 0 or path[-1] != root.data[2]:
+        cur_path.append(root.data[2])
+
+    # leaf node
+    if len(root.eout) == 0:
+        yield tuple(cur_path)
+    # empty node
+    elif len(path) > 1 and len(root.data[2]) == 0:
+        yield tuple(cur_path)
+
+    if len(root.data[2]) > 0:
+        new_path = cur_path
+    else:
+        new_path = [root.data[2]]
+        done.add(root)
+
+    for node in root.eout:
+        if node not in done:
+            yield from get_paths(node, tuple(new_path), done)
+
+# Build an edit graph for each trace and find "meaningful" (to be defined)
+# edits. Return a dictionary of edits and their frequencies, and also
+# submissions and queries in [traces].
+def get_edits_from_traces(traces):
+    # Helper function to remove trailing punctuation from lines. This is a
+    # rather ugly performance-boosting hack.
+    def remove_punct(line):
+        if line and line[-1].type in ('COMMA', 'PERIOD', 'SEMI', 'FROM'):
+            return line[:-1]
+        return line
+
+    # Return values: counts for observed edits, lines, submissions and queries.
+    edits = collections.Counter()
+    lines = collections.Counter()
+    submissions = collections.Counter()
+    queries = collections.Counter()
+
+    for trace in traces:
+        try:
+            actions = parse(trace)
+        except:
+            continue
+        nodes, trace_submissions, trace_queries = edit_graph(actions)
+
+        # Update the submissions/queries counters; rename variables first to
+        # remove trivial differences.
+        for submission in trace_submissions:
+            tokens = tokenize(submission)
+            rename_vars(tokens)
+            code = stringify(tokens)
+            submissions[code] += 1
+
+        for query in trace_queries:
+            tokens = tokenize(query)
+            rename_vars(tokens)
+            code = stringify(tokens)
+            queries[code] += 1
+
+        # Get edits.
+        for path in get_paths(nodes[0]):
+            for i in range(len(path)):
+                start = list(remove_punct(path[i]))
+                var_names = rename_vars(start)
+                start_t = tuple(start)
+
+                for j in range(len(path[i+1:])):
+                    end = list(remove_punct(path[i+1+j]))
+                    rename_vars(end, var_names)
+                    end_t = tuple(end)
+
+                    if start_t != end_t:
+                        edit = (start_t, end_t)
+                        edits[edit] += 1
+                        lines[start_t] += 1
+
+    # Discard rarely occurring edits. XXX only for testing
+    singletons = [edit for edit in edits if edits[edit] < 2]
+    for edit in singletons:
+        lines[edit[0]] -= edits[edit]
+        del edits[edit]
+
+    # Get the probability of each edit given its 'before' line.
+    for before, after in edits:
+        edits[(before, after)] /= lines[before]
+
+    # Normalize line frequencies.
+    if len(lines) > 0:
+        lines_max = max(lines.values())
+        lines = {line: count/lines_max for line, count in lines.items()}
+
+    return edits, lines, submissions, queries
+
+def classify_edits(edits):
+    inserts = {}
+    removes = {}
+    changes = {}
+    for (before, after), cost in edits.items():
+        if after and not before:
+            inserts[after] = cost
+        elif before and not after:
+            removes[before] = cost
+        else:
+            changes[(before, after)] = cost
+    return inserts, removes, changes
+
+# Simplify an edit graph with given nodes: remove empty leaf nodes and other
+# fluff. The function is called recursively until no more changes are done.
+def clean_graph(nodes):
+    changed = False
+
+    # A
+    # | --> A (when B is an empty leaf)
+    # B
+    for node in nodes:
+        if len(node.eout) == 0 and len(node.ein) == 1 and len(node.data[2]) == 0:
+            parent = node.ein[0]
+            parent.eout.remove(node)
+            nodes.remove(node)
+            changed = True
+            break
+
+    # A
+    # | --> A
+    # A
+    for node in nodes:
+        if len(node.ein) == 1:
+            parent = node.ein[0]
+            if len(parent.eout) == 1 and node.data[2] == parent.data[2]:
+                parent.eout = node.eout
+                for child in parent.eout:
+                    child.ein = [parent if node == node else node for node in child.ein]
+                nodes.remove(node)
+                changed = True
+                break
+
+    # A       A
+    # |\      |
+    # | C --> |   (when C is empty)
+    # |/      |
+    # B       B
+    for node in nodes:
+        if len(node.data[2]) == 0 and len(node.ein) == 1 and len(node.eout) == 1:
+            parent = node.ein[0]
+            child = node.eout[0]
+            if len(parent.eout) == 2 and len(child.ein) == 2:
+                parent.eout = [n for n in parent.eout if n != node]
+                child.ein = [n for n in child.ein if n != node]
+                nodes.remove(node)
+                changed = True
+                break
+
+    # A
+    # |
+    # C --> A
+    # |
+    # A
+    for node in nodes:
+        if len(node.data[2]) == 0 and len(node.ein) == 1 and len(node.eout) == 1:
+            parent = node.ein[0]
+            child = node.eout[0]
+            if len(parent.eout) == 1 and len(child.ein) == 1 and parent.data[2] == child.data[2]:
+                parent.eout = [child]
+                child.ein = [parent]
+                nodes.remove(node)
+                changed = True
+                break
+
+    if changed:
+        # go again until nothing changes
+        clean_graph(nodes)
+    else:
+        # compact node ranks
+        ranks = set([node.data[0] for node in nodes])
+        missing = set(range(1,max(ranks)+1)) - ranks
+
+        for node in nodes:
+            diff = 0
+            for rank in sorted(missing):
+                if rank >= node.data[0]:
+                    break
+                diff += 1
+            node.data[0] -= diff
diff --git a/monkey/graph.py b/monkey/graph.py
new file mode 100644
index 0000000..5bf78ec
--- /dev/null
+++ b/monkey/graph.py
@@ -0,0 +1,67 @@
+#!/usr/bin/python3
+
+class Node(object):
+    def __init__(self, data):
+        self.data = data
+        self.ein = []
+        self.eout = []
+
+    # (Re-)insert a child node [target] to [self] at index [idx] (or as the
+    # rightmost child if index is not given). Also append [self] to the list of
+    # parents of [target].
+    def add_out(self, target, idx=None):
+        if target in self.eout:
+            self.eout.remove(target)
+        if idx is None:
+            self.eout.append(target)
+        else:
+            self.eout.insert(idx, target)
+        if self not in target.ein:
+            target.ein.append(self)
+        return target
+
+    def __repr__(self):
+        return str(self.data)
+
+    def __lt__(self, other):
+        return self.data < other.data
+
+# Print the edit graph containing [nodes] in graphviz dot format. The [label]
+# and [pos] functions determine node labels and coordinates (x,y), and the
+# [node_attr] and [edge_attr] functions specify additional attributes for each
+# node and edge. To actually use the coordinates returned by [pos], generate
+# the image using neato -n1.
+def graphviz(nodes, label=str, pos=None, node_attr=None, edge_attr=None):
+    # Generate node descriptions.
+    node_str = ''
+    node_id = {}
+    for node in nodes:
+        node_id[node] = len(node_id)
+        node_str += '\t{} [label="{}"'.format(node_id[node], label(node).replace('"', '\\"'))
+        if pos:
+            node_str += ', ' + 'pos="{},{}"'.format(*pos(node))
+        if node_attr:
+            node_str += ', ' + node_attr(node)
+        node_str += '];\n'
+
+    # Generate edge descriptions (breadth-first).
+    edge_str = ''
+    for node in nodes:
+        a = node_id[node]
+        for child in node.eout:
+            b = node_id[child]
+            edge_str += '\t{} -> {}'.format(a, b)
+            if edge_attr:
+                edge_str += ' [' + edge_attr(node, child) + ']'
+            edge_str += ';\n'
+
+    output = 'digraph G {\n'
+    output += '\tordering="out";\n'
+    output += '\tnode [shape="box", margin="0.05,0", fontname="sans", fontsize=13.0];\n'
+    output += '\n'
+    output += node_str
+    output += '\n'
+    output += edge_str
+    output += '}\n'
+
+    return output
diff --git a/monkey/monkey.py b/monkey/monkey.py
new file mode 100755
index 0000000..42c81f4
--- /dev/null
+++ b/monkey/monkey.py
@@ -0,0 +1,313 @@
+#!/usr/bin/python3
+
+import collections
+import math
+import pickle
+import sys
+import time
+
+from termcolor import colored
+
+import db
+from action import parse
+from edits import classify_edits, clean_graph, edit_graph, get_edits_from_traces
+from graph import Node, graphviz
+from prolog.engine import PrologEngine
+from prolog.util import compose, decompose, map_vars, rename_vars, stringify
+from util import PQueue, Token, indent
+
+# score a program (a list of lines) according to lines distribution
+def score(program, lines):
+    result = 1
+    for line in program:
+        line_normal = list(line)
+        rename_vars(line_normal)
+        line_normal = tuple(line_normal)
+        result *= lines.get(line_normal, 0.01)
+
+    if len(program) == 0 or result == 0:
+        return 0.01
+    return math.pow(result, 1/len(program))
+
+# find a sequence of edits that fixes [code]
+def fix(name, code, edits, timeout=30, debug=False):
+    todo = PQueue()  # priority queue of candidate solutions
+    done = set()     # set of already-analyzed solutions
+
+    # Add a new candidate solution ([lines]+[rules]) to the priority queue.
+    # This solution is generated by applying [step] with [cost] to [prev] task.
+    def add_task(lines, rules, prev=None, step=None, cost=None):
+        if prev is None:
+            path = ()
+            path_cost = 1.0
+        else:
+            path = tuple(list(prev[1]) + [step])
+            path_cost = prev[2] * cost
+        todo.push(((tuple(lines), tuple(rules)), path, path_cost), -path_cost)
+
+    lines, rules = decompose(code)
+    add_task(lines, rules)
+
+    inserts, removes, changes = classify_edits(edits)
+    start_time = time.monotonic()
+    n_tested = 0
+    while True:
+        total_time = time.monotonic() - start_time
+        if total_time > timeout:
+            break
+
+        task = todo.pop()
+        if task == None:
+            break
+
+        (lines, rules), path, path_cost = task
+        code = compose(lines, rules)
+        if code in done:
+            continue
+        done.add(code)
+
+        if debug:
+            print('Cost {:.12f}'.format(path_cost))
+            for line, (before, after) in path:
+                print('line ' + str(line) + ':\t' + stringify(before) + ' → ' + stringify(after))
+
+        # if the code is correct, we are done
+        try:
+            if test(name, code):
+                return code, path, total_time, n_tested
+        except:
+            pass
+        n_tested += 1
+
+        # otherwise generate new solutions
+        rule_no = 0
+        for start, end in rules:
+            rule = lines[start:end]
+            rule_tokens = [t for line in rule for t in line]
+
+            for line_idx in range(start, end):
+                line = lines[line_idx]
+
+                line_normal = list(line)
+                rename_vars(line_normal)
+                line_normal = tuple(line_normal)
+
+                seen = False
+                for (before, after), cost in changes.items():
+                    if line_normal == before:
+                        seen = True
+                        mapping = map_vars(before, after, line, rule_tokens)
+                        after_real = tuple([t if t.type != 'VARIABLE' else Token('VARIABLE', mapping[t.val]) for t in after])
+                        new_lines = lines[:line_idx] + (after_real,) + lines[line_idx+1:]
+                        new_step = ((rule_no, line_idx-start), (tuple(line), after_real))
+
+                        add_task(new_lines, rules, prev=task, step=new_step, cost=cost)
+
+                # if nothing could be done with this line, try removing it
+                # (maybe try removing in any case?)
+                if line_normal in removes.keys() or not seen:
+                    new_lines = lines[:line_idx] + lines[line_idx+1:]
+                    new_rules = []
+                    for old_start, old_end in rules:
+                        new_start, new_end = (old_start - (0 if old_start <= line_idx else 1),
+                                              old_end - (0 if old_end <= line_idx else 1))
+                        if new_end > new_start:
+                            new_rules.append((new_start, new_end))
+                    new_step = ((rule_no, line_idx-start), (tuple(line), ()))
+                    new_cost = removes[line_normal] if line_normal in removes.keys() else 0.9
+
+                    add_task(new_lines, new_rules, prev=task, step=new_step, cost=new_cost)
+
+            # try adding a line to this rule… would need to distinguish between
+            # head/body lines in transforms
+            for after, cost in inserts.items():
+                mapping = map_vars([], after, [], rule_tokens)
+                after_real = [t if t.type != 'VARIABLE' else Token('VARIABLE', mapping[t.val]) for t in after]
+                after_real = tuple(after_real)
+                new_lines = lines[:end] + (after_real,) + lines[end:]
+                new_rules = []
+                for old_start, old_end in rules:
+                    new_rules.append((old_start + (0 if old_start < end else 1),
+                                      old_end + (0 if old_end < end else 1)))
+                new_step = ((rule_no, end-start), ((), after_real))
+
+                add_task(new_lines, new_rules, prev=task, step=new_step, cost=cost)
+            rule_no += 1
+
+        # try adding a new fact
+        if len(rules) < 2:
+            for after, cost in inserts.items():
+                new_lines = lines + (after,)
+                new_rules = rules + (((len(lines), len(lines)+1)),)
+                new_step = ((len(new_rules)-1, 0), (tuple(), tuple(after)))
+
+                add_task(new_lines, new_rules, prev=task, step=new_step, cost=cost)
+
+    return '', [], total_time, n_tested
+
+def print_hint(solution, steps, fix_time, n_tested):
+    if solution:
+        print(colored('Hint found! Tested {} programs in {:.1f} s.'.format(n_tested, fix_time), 'green'))
+        print(colored(' Edits', 'blue'))
+        for line, (before, after) in steps:
+            print('  {}:\t{} → {}'.format(line, stringify(before), stringify(after)))
+        print(colored(' Final version', 'blue'))
+        print(indent(compose(*decompose(solution)), 2))
+    else:
+        print(colored('Hint not found! Tested {} programs in {:.1f} s.'.format(n_tested, fix_time), 'red'))
+
+# Find official solutions to all problems.
+def init_problems():
+    names = {}
+    codes = {}
+    libraries = {}
+
+    pids = db.get_problem_ids()
+    for pid in pids:
+        names[pid], codes[pid], libraries[pid] = db.get_problem(pid)
+
+    return names, codes, libraries
+
+# Submit code to Prolog server for testing.
+def test(name, code):
+    # TODO also load fact library and solved predicates
+    engine = PrologEngine(code=code)
+    result = engine.ask("run_tests({}, '{}')".format(name, engine.id))
+    engine.destroy()
+    return result['event'] == 'success'
+
+if __name__ == '__main__':
+    # Get problem id from commandline.
+    if len(sys.argv) < 2:
+        print('usage: ' + sys.argv[0] + ' <pid>')
+        sys.exit(1)
+    pid = int(sys.argv[1])
+
+    names, codes, libraries = init_problems()
+
+    # Analyze traces for this problem to get edits, submissions and queries.
+    traces = db.get_traces(pid)
+    edits, lines, submissions, queries = get_edits_from_traces(traces.values())
+
+    # Find incorrect submissions.
+    incorrect = []
+    for submission, count in sorted(submissions.items()):
+        if not test(names[pid], submission):
+            # This incorrect submission appeared in [count] attempts.
+            incorrect += [submission]*count
+
+    # XXX only for testing
+    try:
+        done = pickle.load(open('status-'+str(pid)+'.pickle', 'rb'))
+    except:
+        done = []
+
+    # test fix() on incorrect student submissions
+    if len(sys.argv) >= 3 and sys.argv[2] == 'test':
+        timeout = int(sys.argv[3]) if len(sys.argv) >= 4 else 10
+
+        print('Fixing {}/{} programs (timeout={})…'.format(
+            len([p for p in incorrect if p not in done]), len(incorrect), timeout))
+
+        for i, program in enumerate(incorrect):
+            if program in done:
+                continue
+            print(colored('Analyzing program {0}/{1}…'.format(i+1, len(incorrect)), 'yellow'))
+            print(indent(compose(*decompose(program)), 2))
+
+            solution, steps, fix_time, n_tested = fix(names[pid], program, edits, timeout=timeout)
+            if solution:
+                done.append(program)
+            print_hint(solution, steps, fix_time, n_tested)
+            print()
+
+            pickle.dump(done, open('status-'+str(pid)+'.pickle', 'wb'))
+
+        print('Found hints for ' + str(len(done)) + ' of ' + str(len(incorrect)) + ' incorrect programs')
+
+    # print info for this problem
+    elif len(sys.argv) >= 3 and sys.argv[2] == 'info':
+        # with no additional arguments, print some stats
+        if len(sys.argv) == 3:
+            print('Problem {} ({}): {} edits in {} traces, fixed {}/{} ({}/{} unique)'.format(
+                pid, colored(names[pid], 'yellow'),
+                colored(str(len(edits)), 'yellow'), colored(str(len(traces)), 'yellow'),
+                colored(str(len([p for p in incorrect if p in done])), 'yellow'),
+                colored(str(len(incorrect)), 'yellow'),
+                colored(str(len(set(done))), 'yellow'),
+                colored(str(len(set(incorrect))), 'yellow')))
+        else:
+            if sys.argv[3] == 'users':
+                print(' '.join([str(uid) for (pid, uid) in sorted(traces.keys())]))
+            # print all observed edits and their costs
+            elif sys.argv[3] == 'edits':
+                inserts, removes, changes = classify_edits(edits)
+                print('Inserts')
+                for after, cost in sorted(inserts.items(), key=lambda x: x[1]):
+                    print(' {:.2f}\t{}'.format(cost, stringify(after)))
+                print('Removes')
+                for before, cost in sorted(removes.items(), key=lambda x: x[1]):
+                    print(' {:.2f}\t{}'.format(cost, stringify(before)))
+                print('Changes')
+                for (before, after), cost in sorted(changes.items(), key=lambda x: x[1]):
+                    print(' {:.2f}\t{} → {}'.format(cost,
+                                                   stringify(before if before else [('INVALID', 'ε')]),
+                                                   stringify(after if after else [('INVALID', 'ε')])))
+            # print all student submissions not (yet) corrected
+            elif sys.argv[3] == 'unsolved':
+                for p in sorted(set(incorrect)):
+                    if p in done:
+                        continue
+                    print(indent(compose(*decompose(p)), 2))
+                    print()
+            # print all student queries and their counts
+            elif sys.argv[3] == 'queries':
+                for query, count in queries.most_common():
+                    print('  ' + str(count) + '\t' + query)
+
+    # Print the edit graph in graphviz dot syntax.
+    elif len(sys.argv) == 4 and sys.argv[2] == 'graph':
+        uid = int(sys.argv[3])
+        actions = parse(traces[(pid, uid)])
+
+        nodes, submissions, queries = edit_graph(actions)
+
+        def position(node):
+            return (node.data[1]*150, node.data[0]*-60)
+
+        def label(node):
+            return stringify(node.data[2])
+
+        def node_attr(node):
+            if node.ein and node.data[2] == node.ein[0].data[2]:
+                return 'color="gray", shape="point"'
+            return ''
+
+        def edge_attr(a, b):
+            if a.data[2] == b.data[2]:
+                return 'arrowhead="none"'
+            return ''
+
+        graphviz_str = graphviz(nodes, pos=position, label=label,
+                                node_attr=node_attr, edge_attr=edge_attr)
+        print(graphviz_str)
+
+    # run interactive loop
+    else:
+        while True:
+            # read the program from stdin
+            print('Enter program, end with empty line:')
+            code = ''
+            try:
+                while True:
+                    line = input()
+                    if not line:
+                        break
+                    code += line + '\n'
+            except EOFError:
+                break
+
+            # try finding a fix
+            print(colored('Analyzing program…', 'yellow'))
+            solution, steps, fix_time, n_tested = fix(names[pid], code, edits, debug=True)
+            print_hint(solution, steps, fix_time, n_tested)
diff --git a/prolog/engine.py b/monkey/prolog/engine.py
index dff577c..dff577c 100644
--- a/prolog/engine.py
+++ b/monkey/prolog/engine.py
diff --git a/prolog/lexer.py b/monkey/prolog/lexer.py
index 971e8a6..971e8a6 100644
--- a/prolog/lexer.py
+++ b/monkey/prolog/lexer.py
diff --git a/prolog/util.py b/monkey/prolog/util.py
index 0ab3c8b..0ab3c8b 100644
--- a/prolog/util.py
+++ b/monkey/prolog/util.py
diff --git a/monkey/util.py b/monkey/util.py
new file mode 100644
index 0000000..b8be2bb
--- /dev/null
+++ b/monkey/util.py
@@ -0,0 +1,81 @@
+#!/usr/bin/python3
+
+from collections import namedtuple
+from heapq import heappush, heappop
+import itertools
+
+# A simple priority queue based on the heapq class.
+class PQueue(object):
+    REMOVED = '<removed-task>'      # placeholder for a removed task
+
+    def __init__(self):
+        self.pq = []                         # list of entries arranged in a heap
+        self.entry_finder = {}               # mapping of tasks to entries
+        self.counter = itertools.count()     # unique sequence count
+        self.size = 0
+
+    def push(self, task, priority=0):
+        'Add a new task or update the priority of an existing task'
+        if task in self.entry_finder:
+            self.remove(task)
+        else:
+            self.size += 1
+        entry = [priority, next(self.counter), task]
+        self.entry_finder[task] = entry
+        heappush(self.pq, entry)
+
+    def remove(self, task):
+        'Mark an existing task as REMOVED.  Raise KeyError if not found.'
+        entry = self.entry_finder.pop(task)
+        entry[-1] = self.REMOVED
+        self.size -= 1
+
+    def pop(self):
+        'Remove and return the lowest priority task. Raise KeyError if empty.'
+        while self.pq:
+            priority, count, task = heappop(self.pq)
+            if task is not self.REMOVED:
+                del self.entry_finder[task]
+                self.size -= 1
+                return task
+        return None
+
+    def __len__(self):
+        return self.size
+
+# Stores a token's type and value, and optionally the position of the first
+# character in the lexed stream.
+class Token(namedtuple('Token', ['type', 'val', 'pos'])):
+    __slots__ = ()
+
+    # Custom constructor to support default parameters.
+    def __new__(cls, type, val='', pos=None):
+        return super(Token, cls).__new__(cls, type, val, pos)
+
+    def __str__(self):
+        return self.val
+
+    # Ignore position when comparing tokens. There is probably a cleaner way of
+    # doing these.
+    __eq__ = lambda x, y: x[0] == y[0] and x[1] == y[1]
+    __ne__ = lambda x, y: x[0] != y[0] or x[1] != y[1]
+    __lt__ = lambda x, y: tuple.__lt__(x[0:2], y[0:2])
+    __le__ = lambda x, y: tuple.__le__(x[0:2], y[0:2])
+    __ge__ = lambda x, y: tuple.__ge__(x[0:2], y[0:2])
+    __gt__ = lambda x, y: tuple.__gt__(x[0:2], y[0:2])
+
+    # Only hash token's value (we don't care about position, and types are
+    # determined by values).
+    def __hash__(self):
+        return hash(self[1])
+
+# Return [n]th line in [text].
+def get_line(text, n):
+    lines = text.split('\n')
+    if n >= 0 and n < len(lines):
+        return lines[n]
+    return None
+
+# Indent each line in [text] by [indent] spaces.
+def indent(text, indent=2):
+    return '\n'.join([' '*indent+line for line in text.split('\n')])