示例#1
0
def main():
    from sys import stdin
    readline = stdin.readline

    from builtins import max, min, range

    INF = 10**6

    H, W = map(int, readline().split())
    Ch, Cw = map(lambda x: int(x) - 1, readline().split())
    Dh, Dw = map(lambda x: int(x) - 1, readline().split())
    S = [readline()[:-1] for _ in range(H)]

    t = [[INF] * W for _ in range(H)]
    for h in range(H):
        th = t[h]
        Sh = S[h]
        for w in range(W):
            if Sh[w] == '#':
                th[w] = -1

    t[Ch][Cw] = 0
    q = deque([(Ch, Cw)])
    warp_count = 0
    warpq = []
    while q:
        while q:
            warpq.append(q[0])
            h, w = q.popleft()
            if h - 1 >= 0 and t[h - 1][w] > warp_count:
                q.append((h - 1, w))
                t[h - 1][w] = warp_count
            if h + 1 < H and t[h + 1][w] > warp_count:
                q.append((h + 1, w))
                t[h + 1][w] = warp_count
            if w - 1 >= 0 and t[h][w - 1] > warp_count:
                q.append((h, w - 1))
                t[h][w - 1] = warp_count
            if w + 1 < W and t[h][w + 1] > warp_count:
                q.append((h, w + 1))
                t[h][w + 1] = warp_count

        if t[Dh][Dw] != INF:
            break

        warp_count += 1
        for h, w in warpq:
            for i in range(max(0, h - 2), min(H, h + 3)):
                ti = t[i]
                for j in range(max(0, w - 2), min(W, w + 3)):
                    if ti[j] > warp_count:
                        ti[j] = warp_count
                        q.append((i, j))
        warpq.clear()

    if t[Dh][Dw] == INF:
        print(-1)
    else:
        print(t[Dh][Dw])
示例#2
0
    def numpy_max_pool_nd_stride_pad(input,
                                     ws,
                                     ignore_border=True,
                                     stride=None,
                                     pad=None,
                                     mode="max"):
        assert ignore_border
        nd = len(ws)
        if pad is None:
            pad = (0, ) * nd
        if stride is None:
            stride = (0, ) * nd
        assert len(pad) == len(ws) == len(stride)
        assert all(ws[i] > pad[i] for i in range(nd))

        def pad_img(x):
            # initialize padded input
            y = np.zeros(
                x.shape[0:-nd] + tuple(x.shape[-nd + i] + pad[i] * 2
                                       for i in range(nd)),
                dtype=x.dtype,
            )
            # place the unpadded input in the center
            block = (slice(None), ) * (len(x.shape) - nd) + tuple(
                slice(pad[i], x.shape[-nd + i] + pad[i]) for i in range(nd))
            y[block] = x
            return y

        pad_img_shp = list(input.shape[:-nd])
        out_shp = list(input.shape[:-nd])
        for i in range(nd):
            padded_size = input.shape[-nd + i] + 2 * pad[i]
            pad_img_shp.append(padded_size)
            out_shp.append((padded_size - ws[i]) // stride[i] + 1)
        output_val = np.zeros(out_shp)
        padded_input = pad_img(input)
        func = np.max
        if mode == "sum":
            func = np.sum
        elif mode != "max":
            func = np.average
        inc_pad = mode == "average_inc_pad"

        for l in np.ndindex(*input.shape[:-nd]):
            for r in np.ndindex(*output_val.shape[-nd:]):
                region = []
                for i in range(nd):
                    r_stride = r[i] * stride[i]
                    r_end = builtins.min(r_stride + ws[i],
                                         pad_img_shp[-nd + i])
                    if not inc_pad:
                        r_stride = builtins.max(r_stride, pad[i])
                        r_end = builtins.min(r_end,
                                             input.shape[-nd + i] + pad[i])
                    region.append(slice(r_stride, r_end))
                patch = padded_input[l][region]
                output_val[l][r] = func(patch)
        return output_val
示例#3
0
def sync_node_time(cluster):
    hosts = C_Host.objects.filter(
        Q(project_id=cluster.id) & ~Q(name='localhost') & ~Q(name='127.0.0.1') & ~Q(name='::1'))
    data = []
    times = []
    result = {
        'success':True,
        'data':[]
    }
    for host in hosts:
        gmt_date = core.apps.kubeops_api.adhoc.get_host_time(ip=host.ip, port=host.port, username=host.username,
                                                           password=host.password,
                                                           private_key_path=host.private_key_path)
        GMT_FORMAT = '%a %b %d %H:%M:%S CST %Y'
        date = time.strptime(gmt_date, GMT_FORMAT)
        timeStamp = int(time.mktime(date))
        times.append(timeStamp)
        show_time = time.strftime('%Y-%m-%d %H:%M:%S', date)
        time_data = {
            'hostname': host.name,
            'date': show_time,
        }
        data.append(time_data)
    result['data'] = data
    max = builtins.max(times)
    min = builtins.min(times)
    # 如果最大值减最小值超过5分钟 则判断有错
    if (max-min) > 300000:
        result['success'] = False
    return result
示例#4
0
def _set_bottom(device, target):
    mode = DeviceModes(device.mode)
    current_target = device.target

    if mode is DeviceModes.heat_cool:
        bottom = min(target,
                     config.max)  # go up to target, but don't cross max
        # keep the top unless needing to shift up to keep {gap} degree distance
        top = max(current_target.high, bottom + config.gap)
        new_target = (bottom, top)
    elif mode is DeviceModes.heat or mode is DeviceModes.cool:
        new_target = min((target),
                         config.max)  # go up to target, but don't cross max
    else:
        new_target = current_target
    device.target = new_target
示例#5
0
 def __call__(self, *inputvals):
     assert len(inputvals) == len(self.nondata_inputs) + len(
         self.data_inputs)
     nondata_vals = inputvals[0:len(self.nondata_inputs)]
     data_vals = inputvals[len(self.nondata_inputs):]
     feed_dict = dict(zip(self.nondata_inputs, nondata_vals))
     n = data_vals[0].shape[0]
     for v in data_vals[1:]:
         assert v.shape[0] == n
     for i_start in range(0, n, self.batch_size):
         slice_vals = [
             v[i_start:builtins.min(i_start + self.batch_size, n)]
             for v in data_vals
         ]
         for (var, val) in zip(self.data_inputs, slice_vals):
             feed_dict[var] = val
         results = tf.get_default_session().run(self.outputs,
                                                feed_dict=feed_dict)
         if i_start == 0:
             sum_results = results
         else:
             for i in range(len(results)):
                 sum_results[i] = sum_results[i] + results[i]
     for i in range(len(results)):
         sum_results[i] = sum_results[i] / n
     return sum_results
示例#6
0
def make_split(datastream, split=0.2):
    cls = count_classes(datastream)
    mx = builtins.min(cls.values())
    n = int(mx * split)
    res = list(cls.keys()) | select(lambda c: datastream | where(lambda x: x[
        "class_name"] == c) | as_list | pshuffle | take(n)) | chain
    return res | select(lambda x: basename(x["filename"])) | as_list
示例#7
0
文件: C.py 项目: nkymriy/AtCoder
def main():
    from builtins import int, map, list, print, min
    from collections import Counter
    import sys
    sys.setrecursionlimit(10**6)

    input = sys.stdin.readline
    input_str = (lambda: input().rstrip())
    input_list = (lambda: input().rstrip().split())
    input_number = (lambda: int(input()))
    input_number_list = (lambda: list(map(int, input_list())))

    alphabet = 'abcdefghijklmnopqrstuvwxyz'
    N = input_number()
    S = [None] * N
    for i in range(N):
        S[i] = list(input_str())

    ans = ""
    for c in alphabet:
        t = 10**9
        for i in range(N):
            t = min(t, int(Counter(S[i])[c]))
        ans += c * t
    print(ans)
示例#8
0
def get_mechanics_nearest(lat, lng, skill):
    result = []
    # Get mechanics list within a boundary of .5 around the point
    sw_lat = lat - 1.5
    sw_lng = lng - 1.5
    ne_lat = lat + 1.5
    ne_lng = lng + 1.5
    mechanics_list = get_mechanics_list(sw_lat, sw_lng, ne_lat, ne_lng, skill)
    if mechanics_list:
        if len(mechanics_list) == 1:
            nearest = [0]
        else:
            array = [(x[1], x[2]) for x in mechanics_list]
            kd_tree = spatial.cKDTree(array)
            nearest = kd_tree.query((lat, lng), min(5, len(array)))[1]
        for i in nearest:
            skills = []
            pk = mechanics_list[int(i)][0]
            user = Location.objects.get(pk=mechanics_list[int(i)][0]).user
            userprofile = user.userprofile
            name = get_name_from_user(user)
            miles = "%.1f" % geopy_vincenty(lat, lng, mechanics_list[int(i)][1], mechanics_list[int(i)][2])
            bio = userprofile.short_bio
            icon = get_icon_url_from_user(user)
            for skill in userprofile.skills.all():
                skills.append(skill.skill)
            rating = userprofile.rating
            ratingcount = userprofile.rating_count
            result.append((pk, name, miles, bio, icon, skills, rating, ratingcount))
    return result
示例#9
0
文件: pdf.py 项目: c-PRIMED/puq
def _get_range(sfunc, min, max):
    " Truncate PDFs with long tails"

    num_tails = int(sfunc.ppf(0) == np.NINF) + int(sfunc.ppf(1) == np.PINF)

    _range = options['pdf']['range']
    if num_tails:
        if num_tails == 2:
            range = [(1.0 - _range)/2, (1.0 + _range)/2]
        else:
            range = [1.0 - _range, _range]

    mmin = sfunc.ppf(0)
    if mmin == np.NINF:
        mmin = sfunc.ppf(range[0])
    mmax = sfunc.ppf(1)
    if mmax == np.PINF:
        mmax = sfunc.ppf(range[1])

    if min is not None:
        min = builtins.max(min, mmin)
    else:
        min = mmin

    if max is not None:
        max = builtins.min(max, mmax)
    else:
        max = mmax

    return min, max
示例#10
0
 def chunk_columns(df, chunk_size):
     for start in range(0, df.shape[1], chunk_size):
         chunk = df.iloc[:,
                         range(
                             start,
                             builtins.min(start + chunk_size, df.shape[1]))]
         yield chunk
示例#11
0
def findOpposite(user, facebookDF):

    suspects = []
    contenders = facebookDF.loc[(facebookDF['userid'] != user) &
                                (facebookDF['likes_received'] > 1000) &
                                (facebookDF['friend_count'] > 1500),
                                ('userid', 'friend_count',
                                 'friendships_initiated', 'likes',
                                 'likes_received', 'age', 'tenure', 'gender')]
    for index, row in contenders.iterrows():
        for indexMe, rowMe in facebookDF.loc[facebookDF['userid'] == user,
                                             ('userid', 'friend_count',
                                              'friendships_initiated', 'likes',
                                              'likes_received', 'age',
                                              'tenure', 'gender')].iterrows():
            suspects.append([rowMe, row, 0])

    for duo in suspects:
        doppelScore = computeDoppelScore(duo[0], duo[1])
        duo[2] = doppelScore

    scoreList = []
    for triplet in suspects:
        scoreList.append(triplet[2])
    minScore = min(scoreList)
    for triplet in suspects:
        if triplet[2] == minScore:
            triplet[0] = {'userid': (triplet[0]['userid'])}
            triplet[1] = {'userid': (triplet[1]['userid'])}
            triplet[2] = {'doppelScore': triplet[2]}
            return triplet
示例#12
0
文件: pdf.py 项目: c-PRIMED/puq
def HPDF(data, min=None, max=None):
    """
    Histogram PDF - initialized with points from a histogram.

    This function creates a PDF from a histogram.  This is useful when some other software has
    generated a PDF from your data.

    :param data: A two dimensional array. The first column is the histogram interval mean,
        and the second column is the probability.  The probability values do not need to be
        normalized.
    :param min: A minimum value for the PDF range. If your histogram has values very close
        to 0, and you know values of 0 are impossible, then you should set the ***min*** parameter.
    :param max: A maximum value for the PDF range.
    :type data: 2D numpy array
    :returns: A PDF object.
    """
    x = data[:, 0]
    y = data[:, 1]
    sp = interpolate.splrep(x, y)
    dx = (x[1] - x[0]) / 2.0
    mmin = x[0] - dx
    mmax = x[-1] + dx
    if min is not None:
        mmin = builtins.max(min, mmin)
    if max is not None:
        mmax = builtins.min(max, mmax)
    x = np.linspace(mmin, mmax, options['pdf']['numpart'])
    y = interpolate.splev(x, sp)
    y[y < 0] = 0     # if the extrapolation goes negative...
    return PDF(x, y)
示例#13
0
    def add(self, start, length):
        assert start >= 0
        assert length > 0
        #print(" ADD [%d+%d -%d) to %s" % (start, length, start+length, self.dump()))
        first_overlap = last_overlap = None
        for i, (s_start, s_length) in enumerate(self._spans):
            #print("  (%d+%d)-> overlap=%s adjacent=%s" % (s_start,s_length, overlap(s_start, s_length, start, length), adjacent(s_start, s_length, start, length)))
            if (overlap(s_start, s_length, start, length)
                    or adjacent(s_start, s_length, start, length)):
                last_overlap = i
                if first_overlap is None:
                    first_overlap = i
                continue
            # no overlap
            if first_overlap is not None:
                break
        #print("  first_overlap", first_overlap, last_overlap)
        if first_overlap is None:
            # no overlap, so just insert the span and sort by starting
            # position.
            self._spans.insert(0, (start, length))
            self._spans.sort()
        else:
            # everything from [first_overlap] to [last_overlap] overlapped
            first_start, first_length = self._spans[first_overlap]
            last_start, last_length = self._spans[last_overlap]
            newspan_start = min(start, first_start)
            newspan_end = max(start + length, last_start + last_length)
            newspan_length = newspan_end - newspan_start
            newspan = (newspan_start, newspan_length)
            self._spans[first_overlap:last_overlap + 1] = [newspan]
        #print("  ADD done: %s" % self.dump())
        self._check()

        return self
def dijkstra(G):
    """
    Dijkstra algorithm for finding shortest path from start position to end.
    """
    srcIdx = G.vex2idx[G.startpos]
    dstIdx = G.vex2idx[G.endpos]

    # build dijkstra
    nodes = list(G.neighbors.keys())
    dist = {node: float('inf') for node in nodes}
    prev = {node: None for node in nodes}
    dist[srcIdx] = 0

    while nodes:
        curNode = min(nodes, key=lambda node: dist[node])
        nodes.remove(curNode)
        if dist[curNode] == float('inf'):
            break

        for neighbor, cost in G.neighbors[curNode]:
            newCost = dist[curNode] + cost
            if newCost < dist[neighbor]:
                dist[neighbor] = newCost
                prev[neighbor] = curNode

    # retrieve path
    path = deque()
    curNode = dstIdx
    while prev[curNode] is not None:
        path.appendleft(G.vertices[curNode])
        curNode = prev[curNode]
    path.appendleft(G.vertices[curNode])
    return list(path)
示例#15
0
def sync_node_time(cluster):
    hosts = C_Host.objects.filter(
        Q(project_id=cluster.id) & ~Q(name='localhost') & ~Q(name='127.0.0.1')
        & ~Q(name='::1'))
    data = []
    times = []
    result = {'success': True, 'data': []}
    for host in hosts:
        ssh_config = SshConfig(host=host.ip,
                               port=host.port,
                               username=host.username,
                               password=host.password,
                               private_key=None)

        ssh_client = SSHClient(ssh_config)
        res = ssh_client.run_cmd('date')
        gmt_date = res[0]
        GMT_FORMAT = '%a %b %d %H:%M:%S CST %Y'
        date = time.strptime(gmt_date, GMT_FORMAT)
        timeStamp = int(time.mktime(date))
        times.append(timeStamp)
        show_time = time.strftime('%Y-%m-%d %H:%M:%S', date)
        time_data = {
            'hostname': host.name,
            'date': show_time,
        }
        data.append(time_data)
    result['data'] = data
    max = builtins.max(times)
    min = builtins.min(times)
    # 如果最大值减最小值超过5分钟 则判断有错
    if (max - min) > 300000:
        result['success'] = False
    return result
示例#16
0
def HPDF(data, min=None, max=None):
    """
    Histogram PDF - initialized with points from a histogram.

    This function creates a PDF from a histogram.  This is useful when some other software has
    generated a PDF from your data.

    :param data: A two dimensional array. The first column is the histogram interval mean,
        and the second column is the probability.  The probability values do not need to be
        normalized.
    :param min: A minimum value for the PDF range. If your histogram has values very close
        to 0, and you know values of 0 are impossible, then you should set the ***min*** parameter.
    :param max: A maximum value for the PDF range.
    :type data: 2D numpy array
    :returns: A PDF object.
    """
    x = data[:, 0]
    y = data[:, 1]
    sp = interpolate.splrep(x, y)
    dx = (x[1] - x[0]) / 2.0
    mmin = x[0] - dx
    mmax = x[-1] + dx
    if min is not None:
        mmin = builtins.max(min, mmin)
    if max is not None:
        mmax = builtins.min(max, mmax)
    x = np.linspace(mmin, mmax, options['pdf']['numpart'])
    y = interpolate.splev(x, sp)
    y[y < 0] = 0  # if the extrapolation goes negative...
    return PDF(x, y)
示例#17
0
    def test_more_hypothesis(self, peers, shares):
        """
        similar to test_unhappy we test that the resulting happiness is
        always either the number of peers or the number of shares
        whichever is smaller.
        """
        # https://hypothesis.readthedocs.io/en/latest/data.html#hypothesis.strategies.sets
        # hypothesis.strategies.sets(elements=None, min_size=None, average_size=None, max_size=None)[source]
        # XXX would be nice to paramaterize these by hypothesis too
        readonly_peers = set()
        peers_to_shares = {}

        places = happiness_upload.share_placement(peers, readonly_peers,
                                                  set(list(shares)),
                                                  peers_to_shares)
        happiness = happiness_upload.calculate_happiness(places)

        # every share should get placed
        assert set(places.keys()) == shares

        # we should only use peers that exist
        assert set(places.values()).issubset(peers)

        # if we have more shares than peers, happiness is at most # of
        # peers; if we have fewer shares than peers happiness is capped at
        # # of peers.
        assert happiness == min(len(peers), len(shares))
示例#18
0
def _get_range(sfunc, min, max):
    " Truncate PDFs with long tails"

    num_tails = int(sfunc.ppf(0) == np.NINF) + int(sfunc.ppf(1) == np.PINF)

    _range = options['pdf']['range']
    if num_tails:
        if num_tails == 2:
            range = [(1.0 - _range) / 2, (1.0 + _range) / 2]
        else:
            range = [1.0 - _range, _range]

    mmin = sfunc.ppf(0)
    if mmin == np.NINF:
        mmin = sfunc.ppf(range[0])
    mmax = sfunc.ppf(1)
    if mmax == np.PINF:
        mmax = sfunc.ppf(range[1])

    if min is not None:
        min = builtins.max(min, mmin)
    else:
        min = mmin

    if max is not None:
        max = builtins.min(max, mmax)
    else:
        max = mmax

    return min, max
示例#19
0
文件: pipe.py 项目: yueyun00/Pipe
def min(iterable, **kwargs):
    warnings.warn(
        "pipe.min is deprecated, use the builtin min() instead.",
        DeprecationWarning,
        stacklevel=4,
    )
    return builtins.min(iterable, **kwargs)
def new_vertex(randvex, nearvex, stepSize):
    dirn = np.array(randvex) - np.array(nearvex)
    length = np.linalg.norm(dirn)
    dirn = (dirn / length) * min(stepSize, length)

    newvex = (nearvex[0] + dirn[0], nearvex[1] + dirn[1], nearvex[2] + dirn[2])
    return newvex
示例#21
0
def drawSet(set, dom):
    seq = [item['value'] for item in set]
    max = builtins.max(seq)
    min = builtins.min(seq)

    height = int(dom.getAttribute("SVG", "height"))

    min = 0 if min > 0 else min

    svg = atlastk.createHTML()

    for i in range(len(set)):
        svg.pushTag("rect")
        svg.putAttribute("x",
                         str(i * 100 / len(set)) + "%")
        svg.putAttribute("y", height - set[i]["value"] * height / max)
        svg.putAttribute("width",
                         str(100 / len(set)) + "%")
        svg.putAttribute("height",
                         str(100 * set[i]["value"] / max) + "%")
        svg.putTagAndValue("title",
                           set[i]["date"] + " : " + str(set[i]["value"]))
        svg.popTag()

    dom.inner("SVG", svg)

    dom.setValue("Text", set[0]["date"] + " - " + set[len(set) - 1]["date"])
示例#22
0
def getKeyRange(m):
    low = 0xFFFFFFFF
    high = 0
    for key in m:
        iKey = int(key)
        low = min(iKey, low)
        high = max(iKey, high)
    return Range(low, high)
 def __init__(self, point_map, original, generated):
     self.generated = generated
     self.original = original
     self.point_map = point_map
     self.width = len(point_map[0])
     self.height = len(point_map)
     self.cell_size = int(min(MAX_X / self.width, MAX_Y / self.height))
     self.offset = self.cell_size / 2
示例#24
0
文件: test_a.py 项目: spack971/none
 def __init__(self, *args: ty.Optional[int]):
     s = builtins.slice(*args)
     start, stop, step = (
         s.start or 0,
         s.stop or builtins.min(s.stop, MAX_RANGE),
         s.step or 1,
     )
     self._it = builtins.iter(builtins.range(start, stop, step))
示例#25
0
def shortest_path(edge, mid_contour, w, h):
    pix_val = []
    half_cols = w // 2
    total_cols = w
    total_rows = h
    #  range goes from halfway through the x direction and
    #  the whole way in the y direction
    for i in range(half_cols):
        for j in range(total_rows):
            g = edge[j][i][1]
            if g == 255:
                pix_val.append([i, j])

    min_distance = float("inf")

    val_x1, val_y1, val_x2, val_y2 = -1, -1, -1, -1

    for coor in pix_val:
        col, row = coor[0], coor[1]
        theta = math.atan2((mid_contour[0] - col), (mid_contour[1] - row))
        for radius in range(int(min(total_rows, total_cols) / 2)):
            new_col = min(mid_contour[0] + math.sin(theta) * radius,
                          total_cols - 1)
            new_row = min(mid_contour[1] + math.cos(theta) * radius,
                          total_rows - 1)
            # print((new_col, new_row))
            g = edge[int(new_row)][int(new_col)][1]
            if g == 255:
                dist = distance(new_col, new_row, col, row)
                if dist < min_distance:
                    val_x1 = col
                    val_y1 = row
                    val_x2 = new_col
                    val_y2 = new_row
                    min_distance = dist

    cv2.circle(edge, (int(val_x1), int(val_y1)), 5, (255, 0, 255), -1)
    cv2.circle(edge, (int(val_x2), int(val_y2)), 5, (255, 0, 255), -1)
    cv2.circle(edge, (int(mid_contour[0]), int(mid_contour[1])), 5,
               (255, 0, 255), -1)

    cv2.imshow("points", edge)
    cv2.waitKey(0)

    return "Shortest path: ", val_x1, val_y1, " to ", val_x2, val_y2, " Distance: ", min_distance
示例#26
0
 def get_best(self, data_name, is_larger_better=True):
     if data_name in self:
         steps, values = zip(*self[data_name].copy())
         if is_larger_better:
             return builtins.max(values)
         else:
             return builtins.min(values)
     else:
         raise ValueError('{0} is not in this History.'.format(data_name))
示例#27
0
文件: audioop.py 项目: Qointum/pypy
def minmax(cp, size):
    _check_params(len(cp), size)

    min_sample, max_sample = 0x7fffffff, -0x80000000
    for sample in _get_samples(cp, size):
        max_sample = builtins.max(sample, max_sample)
        min_sample = builtins.min(sample, min_sample)

    return min_sample, max_sample
示例#28
0
文件: audioop.py 项目: Qointum/pypy
def minmax(cp, size):
    _check_params(len(cp), size)

    min_sample, max_sample = 0x7fffffff, -0x80000000
    for sample in _get_samples(cp, size):
        max_sample = builtins.max(sample, max_sample)
        min_sample = builtins.min(sample, min_sample)

    return min_sample, max_sample
    def on_loss_calculation_end(self, training_context):
        """Returns mixed inputs, pairs of targets, and lambda"""
        train_data = training_context['train_data']
        x = None
        y = None
        x = train_data.value_list[0].copy().detach()  # input
        y = train_data.value_list[1].copy().detach()  # label
        model = training_context['current_model']

        lam = builtins.min(
            builtins.max(np.random.beta(self.alpha, self.alpha), 0.3), 0.7)

        batch_size = int_shape(x)[0]
        index = arange(batch_size)
        index = cast(shuffle(index), 'long')
        this_loss = None
        mixed_x = None
        if get_backend() == 'pytorch':
            mixed_x = lam * x + (1 - lam) * x[index, :]
            pred = model(to_tensor(mixed_x, requires_grad=True))
            y_a, y_b = y, y[index]
            this_loss = lam * self.loss_criterion(pred, y_a.long()) + (
                1 - lam) * self.loss_criterion(pred, y_b.long())
        elif get_backend() == 'tensorflow':
            x1 = tf.gather(x, index, axis=0)
            y1 = tf.gather(y, index, axis=0)
            mixed_x = lam * x + (1 - lam) * x1
            pred = model(to_tensor(mixed_x, requires_grad=True))
            y_a, y_b = y, y1

            this_loss = lam * self.loss_criterion(
                pred, y_a) + (1 - lam) * self.loss_criterion(pred, y_b)

        training_context['current_loss'] = training_context[
            'current_loss'] + this_loss * self.loss_weight
        if training_context['is_collect_data']:
            training_context['losses'].collect(
                'mixup_loss', training_context['steps'],
                float(to_numpy(this_loss * self.loss_weight)))

        if training_context['current_batch'] == 0:
            for item in mixed_x:
                if self.save_path is None and not is_in_colab():
                    item = unnormalize([0.485, 0.456, 0.406],
                                       [0.229, 0.224, 0.225])(to_numpy(item))
                    item = unnormalize(0, 255)(item)
                    array2image(item).save('Results/mixup_{0}.jpg'.format(
                        get_time_suffix()))
                elif self.save_path is not None:
                    item = unnormalize([0.485, 0.456, 0.406],
                                       [0.229, 0.224, 0.225])(to_numpy(item))
                    item = unnormalize(0, 255)(item)
                    array2image(item).save(
                        os.path.join(self.save_path, 'mixup_{0}.jpg'.format(
                            get_time_suffix())))
    def auto_canny(image, sigma=0.33):
        # compute the median of the single channel pixel intensities
        v = np.median(image)

        # apply automatic Canny edge detection using the computed median
        lower = int(max(0, (1.0 - sigma) * v))
        upper = int(min(255, (1.0 + sigma) * v))
        edged = cv2.Canny(image, lower, upper)

        # return the edged image
        return edged
示例#31
0
文件: test_a.py 项目: spack971/none
 def __init__(self, *args: ty.Optional[int], elements: int = 1):
     s = builtins.slice(*args)
     start, stop, step = (
         s.start or 0,
         s.stop or builtins.min(s.stop, MAX_RANGE),
         s.step or 1,
     )
     self._it = builtins.tuple(
         builtins.iter(builtins.range(start, stop, step))
         for _ in builtins.range(elements)
     )
示例#32
0
    def img_op(image: np.ndarray, **kwargs):
        image = np.clip(image, 0.0, 255.0)
        gammamin, gammamax = gamma_range
        avg_pix = image.mean()
        if avg_pix > 220:
            gammamax = builtins.max(gammamin, 1)
        elif avg_pix < 30:
            gammamin = builtins.min(1, gammamax)

        gamma = np.random.choice(np.arange(gammamin, gammamax, 0.01))
        return exposure.adjust_gamma(image / 255.0, gamma) * 255.0
示例#33
0
def overlap(start0, length0, start1, length1):
    # return start2,length2 of the overlapping region, or None
    #  00      00   000   0000  00  00 000  00   00  00      00
    #     11    11   11    11   111 11 11  1111 111 11    11
    left = max(start0, start1)
    right = min(start0 + length0, start1 + length1)
    # if there is overlap, 'left' will be its start, and right-1 will
    # be the end'
    if left < right:
        return (left, right - left)
    return None
示例#34
0
def main():
    from builtins import int, map, list, print, min
    import sys
    sys.setrecursionlimit(10**6)

    input = sys.stdin.readline
    input_list = (lambda: input().rstrip().split())
    input_number = (lambda: int(input()))
    input_number_list = (lambda: list(map(int, input_list())))

    K, N = input_number_list()
    A = input_number_list()
    A.sort()
    ans = A[-1] - A[0]
    for i in range(1, N - 1):
        l = A[i] + (K - A[i + 1])
        r = K - A[i] + A[i - 1]
        ans = min(ans, l, r)
    ans = min(ans, K - A[-1] + A[-2])
    print(ans)
示例#35
0
def min(*args):
    """Override the builtin min function to expand list arguments.

    Arguments:
        *args -- lists of numbers or individual numbers
    """
    fullList = []
    for arg in args:
        if hasattr(arg, 'extend'):
            fullList.extend(arg)
        else:
            fullList.append(arg)
    if not fullList:
        return 0
    return builtins.min(fullList)
示例#36
0
 def __call__(self, *inputvals):
     assert len(inputvals) == len(self.nondata_inputs) + len(self.data_inputs)
     nondata_vals = inputvals[0:len(self.nondata_inputs)]
     data_vals = inputvals[len(self.nondata_inputs):]
     feed_dict = dict(zip(self.nondata_inputs, nondata_vals))
     n = data_vals[0].shape[0]
     for v in data_vals[1:]:
         assert v.shape[0] == n
     for i_start in range(0, n, self.batch_size):
         slice_vals = [v[i_start:builtins.min(i_start + self.batch_size, n)] for v in data_vals]
         for (var, val) in zip(self.data_inputs, slice_vals):
             feed_dict[var] = val
         results = tf.get_default_session().run(self.outputs, feed_dict=feed_dict)
         if i_start == 0:
             sum_results = results
         else:
             for i in range(len(results)):
                 sum_results[i] = sum_results[i] + results[i]
     for i in range(len(results)):
         sum_results[i] = sum_results[i] / n
     return sum_results
示例#37
0
def _nanmin(values, axis=None, skipna=True):
    values, mask, dtype = _get_values(values, skipna, fill_value_typ = '+inf')

    # numpy 1.6.1 workaround in Python 3.x
    if (values.dtype == np.object_
            and sys.version_info[0] >= 3):  # pragma: no cover
        import builtins
        if values.ndim > 1:
            apply_ax = axis if axis is not None else 0
            result = np.apply_along_axis(builtins.min, apply_ax, values)
        else:
            result = builtins.min(values)
    else:
        if ((axis is not None and values.shape[axis] == 0)
                or values.size == 0):
            result = com.ensure_float(values.sum(axis))
            result.fill(np.nan)
        else:
            result = values.min(axis)

    result = _wrap_results(result,dtype)
    return _maybe_null_out(result, axis, mask)
示例#38
0
def min2(*args):
    return builtins.min(filter(lambda x: x is not None, args),
                        default=None)
示例#39
0
文件: pipe.py 项目: safwank/Pipe
def min(iterable, **kwargs):
    return builtins.min(iterable, **kwargs)
示例#40
0
文件: handlers.py 项目: sirex/databot
def min(expr, pos, value, **kwargs):
    return builtins.min(value, **kwargs)
示例#41
0
def min(a, b):
    """Returns the minimum of a and b."""
    return builtins.min(a, b)