func (s *server) internalCrawl(in *pb.CrawlRequest) (*pb.CrawlResponse, error) {
response := pb.CrawlResponse{}
// 文件名
hasher := sha1.New()
hasher.Write([]byte(in.Url))
sha := base64.URLEncoding.EncodeToString(hasher.Sum(nil))
cacheFilename := fmt.Sprintf("%s/page-%s", *pageCacheDir, sha)
metadataFilename := fmt.Sprintf("%s/metadata-%s", *pageCacheDir, sha)
// 检查页面的 metadata 和 page 是否已经被抓取
// 如果有任何异常或者页面过旧，则重新抓取
if _, err := os.Stat(metadataFilename); err == nil {
serializedMetadata, err := ioutil.ReadFile(metadataFilename)
if err == nil {
response.Metadata = &pb.Metadata{}
err = proto.Unmarshal(serializedMetadata, response.Metadata)
if err == nil {
if in.OnlyReturnMetadata {
return &response, nil
}
if time.Now().UnixNano()-response.Metadata.LastCrawlTimestamp < in.CrawlFrequency*int64(time.Millisecond) {
if _, err := os.Stat(cacheFilename); err == nil {
content, err := ioutil.ReadFile(cacheFilename)
if err == nil {
contentReader := bytes.NewReader(content)
reader, err := gzip.NewReader(contentReader)
if err == nil {
if b, err := ioutil.ReadAll(reader); err == nil {
response.Content = string(b)
return &response, nil
}
reader.Close()
}
}
}
}
}
}
}
// 获取 http 连接
tr := &http.Transport{
TLSClientConfig: &tls.Config{InsecureSkipVerify: true},
}
client := &http.Client{
Transport: tr,
}
if in.Timeout > 0 {
client.Timeout = time.Millisecond * time.Duration(in.Timeout)
}
resp, err := client.Get(in.Url)
if err != nil {
return nil, err
}
// 读取页面内容
body, err := ioutil.ReadAll(resp.Body)
resp.Body.Close()
if err != nil {
return nil, err
}
// 充填 response
if !in.OnlyReturnMetadata {
response.Content = string(body)
}
response.IsFreshCrawl = true
// 充填 metadata
response.Metadata = &pb.Metadata{}
response.Metadata.LastCrawlTimestamp = time.Now().UnixNano()
response.Metadata.Length = uint32(len(body))
// 将 page content 写入文件
var buf bytes.Buffer
writer := gzip.NewWriter(&buf)
writer.Write(body)
writer.Close()
if err := ioutil.WriteFile(cacheFilename, buf.Bytes(), 0644); err != nil {
return &response, err
}
// 将 metadata 写入文件
serializedMetadata, err := proto.Marshal(response.Metadata)
if err != nil {
return &response, err
}
if err := ioutil.WriteFile(metadataFilename, serializedMetadata, 0644); err != nil {
return &response, err
}
return &response, nil
}
func (s *server) internalCrawl(in *pb.CrawlRequest) (*pb.CrawlResponse, error) {
response := pb.CrawlResponse{}
// 仅当 method 为 GET 或者 HEAD 时才读取文件缓存
var cacheFilename, metadataFilename string
if in.Method == pb.Method_GET || in.Method == pb.Method_HEAD {
// 生成缓存文件名
hasher := sha1.New()
hasher.Write([]byte(in.Url))
sha := base64.URLEncoding.EncodeToString(hasher.Sum(nil))
cacheFilename = fmt.Sprintf("%s/page-%s", *pageCacheDir, sha)
metadataFilename = fmt.Sprintf("%s/metadata-%s", *pageCacheDir, sha)
// 检查页面的 metadata 和 page 是否已经被抓取
// 如果有任何异常或者页面过旧，则重新抓取
if _, err := os.Stat(metadataFilename); err == nil {
serializedMetadata, err := ioutil.ReadFile(metadataFilename)
if err == nil {
response.Metadata = &pb.Metadata{}
err = proto.Unmarshal(serializedMetadata, response.Metadata)
if err == nil {
if in.OnlyReturnMetadata {
return &response, nil
}
if time.Now().UnixNano()-response.Metadata.LastCrawlTimestamp <
in.RecrawlTtl*int64(time.Millisecond) {
if _, err := os.Stat(cacheFilename); err == nil {
content, err := ioutil.ReadFile(cacheFilename)
if err == nil {
contentReader := bytes.NewReader(content)
reader, err := gzip.NewReader(contentReader)
if err == nil {
if b, err := ioutil.ReadAll(reader); err == nil {
response.Content = string(b)
return &response, nil
}
reader.Close()
}
}
}
}
}
}
}
}
// 获取 http 连接
tr := &http.Transport{
TLSClientConfig: &tls.Config{InsecureSkipVerify: true},
}
client := &http.Client{
Transport: tr,
}
if in.Timeout > 0 {
client.Timeout = time.Millisecond * time.Duration(in.Timeout)
}
// 根据不同的 method 类型，分别调用不同 HTTP 方法
var resp *http.Response
var err error
if in.Method == pb.Method_GET {
resp, err = client.Get(in.Url)
} else if in.Method == pb.Method_HEAD {
resp, err = client.Head(in.Url)
} else if in.Method == pb.Method_POST {
buff := bytes.NewBufferString(in.PostBody)
resp, err = client.Post(in.Url, in.BodyType, buff)
} else if in.Method == pb.Method_POSTFORM {
formKeys := url.Values{}
for _, kv := range in.FormValues {
formKeys.Add(kv.Key, kv.Value)
}
resp, err = client.PostForm(in.Url, formKeys)
}
if err != nil {
return nil, err
}
// 只有当 method 不为 HEAD 时才读取页面内容
var body []byte
if in.Method != pb.Method_HEAD {
// 读取页面内容
var err error
body, err = ioutil.ReadAll(resp.Body)
resp.Body.Close()
if err != nil {
return nil, err
}
// 充填 response
if !in.OnlyReturnMetadata {
response.Content = string(body)
}
response.IsFreshCrawl = true
}
// 充填 metadata
response.Metadata = &pb.Metadata{}
response.Metadata.LastCrawlTimestamp = time.Now().UnixNano()
response.Metadata.Length = uint32(len(body))
for key, vs := range resp.Header {
for _, v := range vs {
response.Metadata.Header = append(response.Metadata.Header, &pb.KV{
Key: key,
Value: v,
})
}
}
response.Metadata.Status = resp.Status
response.Metadata.StatusCode = int32(resp.StatusCode)
// 仅当 method 为 GET 才将页面内容写入文件
if in.Method == pb.Method_GET {
var buf bytes.Buffer
writer := gzip.NewWriter(&buf)
writer.Write(body)
writer.Close()
if err := ioutil.WriteFile(cacheFilename, buf.Bytes(), 0644); err != nil {
return &response, err
}
}
// 仅当 method 为 GET 或者 HEAD 时才将 metadata 写入文件
if in.Method == pb.Method_GET || in.Method == pb.Method_HEAD {
serializedMetadata, err := proto.Marshal(response.Metadata)
if err != nil {
return &response, err
}
if err := ioutil.WriteFile(metadataFilename, serializedMetadata, 0644); err != nil {
return &response, err
}
}
return &response, nil
}