def upsert(tr: WebElement, fwVer: str): global prevTrail, driver try: trTxt = tr.text relDate = guessDate(trTxt) downLinks = tr / csss / 'td:nth-child(7) div:nth-child(1) > a' fileUrls = [_.get_attribute('data-filelink') for _ in downLinks] fileUrls = [_ for _ in fileUrls] fileUrl = '\n'.join(_ for _ in fileUrls) trailStr = str(prevTrail + [idx]) pageUrl = driver.current_url model = waitText('div.container:nth-child(7) > div > div > h2') prodName = waitText( 'div.container:nth-child(7) > div:nth-child(1) > div:nth-child(1) > p:nth-child(2)' ) sql( "INSERT OR REPLACE INTO TFiles(" " model, prod_name, fw_ver, rel_date, " " page_url, file_url, tree_trail) VALUES" "(:model, :prodName, :fwVer,:relDate, " ":pageUrl, :fileUrl, :trailStr)", glocals()) ulog('UPSERT "%(model)s", "%(fwVer)s", "%(relDate)s", ' ' "%(fileUrl)s", %(trailStr)s ' % glocals()) except Exception as ex: ipdb.set_trace() traceback.print_exc()
def fileWalker(): global driver,prevTrail try: modelName = waitText('h1 strong.model') modelRev = waitText('h1') revName = modelRev.split(modelName)[-1].strip() tabbtn = cssWithText('ul.row li a', 'Firmware') if not tabbtn: ulog('no firmware download for "%s"'%modelName) return tabbtn.click() pageUrl=driver.current_url tables=getElems('#content_firmware table') #waitUntil( lambda:ulog('is_displayed()=%s'%[_.is_displayed() for # _ in tables])>=0 ) startIdx = getStartIdx() numTables=len(tables) for idx in range(startIdx,numTables): table=tables[idx] if not table.is_displayed(): continue ulog('trail=%s'%(prevTrail+[idx])) basicInfo=table.find_element_by_css_selector('tr.basic-info').text fileName=basicInfo.splitlines()[0].strip() fwVer = fileName.split('_')[-1].strip() fwDate=guessDate(basicInfo) fileSize=guessFileSize(basicInfo) fileLink=table.find_element_by_css_selector('a') fileUrl=fileLink.get_attribute('href') ulog('fileName="%s"'%fileName) fwDesc='\n'.join(_.text for _ in table.find_elements_by_css_selector('tr.more-info')) trailStr=str(prevTrail+[idx]) sql("INSERT OR REPLACE INTO TFiles (model,revision," "fw_date, fw_ver, fw_desc, file_name,file_size, " "page_url,file_url,tree_trail) VALUES" "(:modelName, :revName, " ":fwDate,:fwVer,:fwDesc,:fileName,:fileSize," ":pageUrl,:fileUrl,:trailStr)",locals()) ulog('UPSERT "%(modelName)s", "%(revName)s", "%(fwDate)s", ' ' "%(fileName)s", %(fileSize)s,%(fileUrl)s'%locals()) return except Exception as ex: ipdb.set_trace() traceback.print_exc() driver.save_screenshot(getScriptName()+'_'+getFuncName()+'_exc.png')
def pageWalker(): global prevTrail, driver CSS=driver.find_elements_by_css_selector try: startIdx = getStartIdx() startPage = startIdx+1 curPage = 1 idx = curPage-1 while idx != startPage-1: ulog('idx=%d,page=%d'%(idx, (idx+1))) pages = getElems('.x-page-com a') def pageNum(p): try: return int(p.text.strip()) except ValueError: pass href = p.get_attribute('href') if not href: return sys.maxsize try: return int(re.search(r'void\((.+)\)', href).group(1)) except Exception as ex: ipdb.set_trace() traceback.print_exc() tarPage = min(pages, key=lambda p: abs(startPage - pageNum(p))) ulog('tarPage=%d'%pageNum(tarPage)) tarPage.click() ulog('tarPage.click()') time.sleep(0.5) retryUntilTrue(lambda:len(CSS('.x-waite'))==1, 16, 0.4 ) uprint('waitCursor shows') retryUntilTrue(lambda:len(CSS('.x-waite'))==0 or CSS('.x-waite')[0].is_displayed()==False, 60, 1 ) uprint('waitCursor disappears') curPage = int(waitText('a.cur')) ulog('curPage=%d'%curPage) idx = curPage-1 for idx in itertools.count(startIdx): ulog('idx=%d,page=%d'%(idx, (idx+1))) prevTrail+=[idx] rowWalker() prevTrail.pop() try: nextPage = waitClickable('.x-next-on') except (NoSuchElementException, TimeoutException): ulog('last page') break nextPage.click() ulog('nextPage.click()') time.sleep(0.5) retryUntilTrue(lambda:len(CSS('.x-waite'))==1, 16, 0.4 ) uprint('waitCursor shows') retryUntilTrue(lambda:len(CSS('.x-waite'))==0 or CSS('.x-waite')[0].is_displayed()==False, 60, 1 ) uprint('waitCursor disappears') except Exception as ex: ipdb.set_trace() traceback.print_exc()
def modelWalker(): global driver, prevTrail, models act=ActionChains(driver) CSSs = driver.find_elements_by_css_selector try: startIdx = getStartIdx() for idx, model in enumerate(models[startIdx:],len(models)): ulog('idx=%s, model="%s"'%(idx,model)) goToUrl(rootUrl) btn=waitClickable('.search-select button') act.move_to_element(btn).click(btn).perform() inp=waitClickable('.input-block-level') act.move_to_element(inp).click(inp).perform() act.send_keys(model + Keys.DOWN + Keys.ENTER).perform() time.sleep(0.1) waitUntil(isReadyState) ulog('url='+driver.current_url) title = waitText('.lightGrayBg > div > div > div > h2') ulog('title='+title) # 'Search by Model Number' or 'No Matches Found' if title.startswith('No Matches Found'): continue prevTrail+=[idx] tabWalker() prevTrail.pop() except Exception as ex: ipdb.set_trace() traceback.print_exc()
def modelWalker(): global driver, prevTrail, models act = ActionChains(driver) CSSs = driver.find_elements_by_css_selector try: startIdx = getStartIdx() for idx, model in enumerate(models[startIdx:], len(models)): ulog('idx=%s, model="%s"' % (idx, model)) goToUrl(rootUrl) btn = waitClickable('.search-select button') act.move_to_element(btn).click(btn).perform() inp = waitClickable('.input-block-level') act.move_to_element(inp).click(inp).perform() act.send_keys(model + Keys.DOWN + Keys.ENTER).perform() time.sleep(0.1) waitUntil(isReadyState) ulog('url=' + driver.current_url) title = waitText('.lightGrayBg > div > div > div > h2') ulog('title=' + title) # 'Search by Model Number' or 'No Matches Found' if title.startswith('No Matches Found'): continue prevTrail += [idx] tabWalker() prevTrail.pop() except Exception as ex: ipdb.set_trace() traceback.print_exc()
def treeWalker(): global startTrail,prevTrail, driver def getNodes(): return getElems('#psbox3 li a' if prevTrail else '#psbox2 li a') nodes=getNodes() numNodes=len(nodes) if startTrail: startIdx=startTrail.pop(0) else: startIdx=0 ulog('startTrail=%s'%startTrail) ulog('prevTrail=%s'%prevTrail) ulog('startIdx=%d'%startIdx) nodeTxts=[getElemText(_) for _ in getNodes()] ulog('nodes=%s (len=%d)'%([(i,_) for i,_ in enumerate(nodeTxts)],numNodes)) for idx in range(startIdx, numNodes): try: crumbs=waitText('#psa_crumbs').replace('\n',' > ') ulog('crumbs=%s'%crumbs) nodeTxt = getElemText(nodes[idx]) nodeId = nodes[idx].get_attribute('id').strip() ulog('nodeId="%s"'%nodeId) if not nodeId: # isLeaf ulog('Click Leaf "%s"'%nodeTxt) clickElem(nodes[idx]) with suppress(StaleElementReferenceException): nodes[idx].click() prevTrail+=[idx] # if not waitUntil(lambda: not driver.find_elements_by_css_selector('#psbox3')): # nodes[idx].click() selectSoftwareType() prevTrail.pop() assert getCurDepthAtTreeWalker()==len(prevTrail) nodes=getNodes() continue with UntilTextChanged('#psbox3'): ulog('Click branch "%s"'%nodeTxt) clickElem(nodes[idx]) prevTrail+=[idx] treeWalker() prevTrail.pop() nodes=getNodes() except Exception as ex: ipdb.set_trace() print(ex); traceback.print_exc() driver.save_screenshot('cisco_treeWalker.png') # end for crumbs=getElems('#psa_crumbs a') ulog('back to "%s"'%getElemText(crumbs[-1])) ulog('prevTrail=%s'%prevTrail) clickElem(crumbs[-1])
def selectCategory(prev_url): global category, prevTrail, searchResultsNotification,driver try: if len(prevTrail)==1: try: waitVisible('.filter-list', 30, 0.4) except TimeoutException: ulog('No search results, url=%s'%driver.current_url) driver.get(prev_url) return searchResultsNotification=waitText('.search-results-notification').strip() # Your search for f returned 4196 results elif len(prevTrail)==2: searchResultsNotification=waitTextChanged('.search-results-notification', searchResultsNotification).strip() # Your search for f returned 67 results ulog('%s'%searchResultsNotification) category = waitText('.accordion-activate a') ulog('category="%s"'%category) cats=getElems('.filter-list a') retryUntilTrue(lambda:ulog('cats=%s'%[(i,_.text)for i,_ in enumerate(cats)])) numCats=len(cats) startIdx = getStartIdx() for idx in range(startIdx, numCats): ulog('click %s,"%s"'%(idx,cats[idx].text)) prevTrail+=[idx] if len(prevTrail)==2: enterElem(cats[idx], selectCategory) else: enterElem(cats[idx], selectProduct) prevTrail.pop() cats = getElems('.filter-list a') if prev_url: driver.get(prev_url) searchResultsNotification=waitTextChanged('.search-results-notification', searchResultsNotification).strip() except Exception as ex: ipdb.set_trace() traceback.print_exc() driver.save_screenshot(getScriptName()+'_'+getFuncName()+'_excep.png')
def upsert(tr:WebElement, fwVer:str): global prevTrail,driver try: trTxt = tr.text relDate = guessDate(trTxt) downLinks = tr/csss/'td:nth-child(7) div:nth-child(1) > a' fileUrls = [_.get_attribute('data-filelink') for _ in downLinks] fileUrls = [_ for _ in fileUrls] fileUrl = '\n'.join(_ for _ in fileUrls) trailStr=str(prevTrail+[idx]) pageUrl = driver.current_url model = waitText('div.container:nth-child(7) > div > div > h2') prodName = waitText('div.container:nth-child(7) > div:nth-child(1) > div:nth-child(1) > p:nth-child(2)') sql("INSERT OR REPLACE INTO TFiles(" " model, prod_name, fw_ver, rel_date, " " page_url, file_url, tree_trail) VALUES" "(:model, :prodName, :fwVer,:relDate, " ":pageUrl, :fileUrl, :trailStr)", glocals()) ulog('UPSERT "%(model)s", "%(fwVer)s", "%(relDate)s", ' ' "%(fileUrl)s", %(trailStr)s '%glocals()) except Exception as ex: ipdb.set_trace() traceback.print_exc()
def walkFile(): global driver, prevTrail try: try: modelName = waitTextChanged('h2#searchResults', None, 5, 1) except TimeoutException: try: modelName = waitText('h2#searchResults', 5, 1) except TimeoutException: return PROC_GIVE_UP ulog('modelName="%s"'%modelName) resultsCount = waitText('#LargeFirmware>p') # try: # resultsCount = waitTextChanged('#LargeFirmware>p', None, 0.5, 0.25) # ulog('waitTextChanged #LargeFirmware>p') # except TimeoutException: # ulog('TimeoutException: #LargeFirmware>p') # resultsCount = waitText('#LargeFirmware>p') # except NoSuchElementException: # ulog('NoSuchElementException: #LargeFirmware>p') # resultsCount = waitText('#LargeFirmware>p') ulog('resutlsCount=%s'%resultsCount) if resultsCount.startswith('No matching'): return numFiles = int(re.search(r'\d+', resultsCount).group(0)) ulog('numFiles=%d'%numFiles) try: waitTextChanged('#LargeFirmware a.navlistsearch', None, 1, 0.5) ulog('waitTextChanged #LargeFirmware a.navlistsearch') except TimeoutException: ulog('TimeoutException: #LargeFirmware a.navlistsearch') pass except NoSuchElementException: ulog('NoSuchElementException #LargeFirmware a.navlistsearch') return if numFiles > 10: ulog('click moreResults because numFiles=%d>10'%numFiles) bMoreResultsClicked=False for _i in range(10): moreResults = waitClickable('#lnkAllDownloadMore') try: moreResults.click() ulog('moreResults.click()') bMoreResultsClicked=True break except WebDriverException: time.sleep(0.5) if not bMoreResultsClicked: raise StaleElementReferenceException() lastFile = driver.find_element_by_css_selector('#LargeFirmware li:nth-child(%d) a.navlistsearch'%numFiles) for _i in range(10): if lastFile.is_displayed(): break time.sleep(0.5) # try: # waitTextChanged('#LargeFirmware li:nth-child(%d)'%numFiles, # None, 1, 0.5) # ulog('waitTextChanged #LargeFirmware li:nth-child(%d)'%numFiles) # except TimeoutException: # ulog('TimeoutException: #LargeFirmware li:nth-child(%d)'%numFiles) # pass # except NoSuchElementException: # ulog('NoSuchElementException #LargeFirmware li:nth-child(%d)'%numFiles) # return # waitClickable('#LargeFirmware li:nth-child(%d) a.navlistsearch' # %numFiles) files = getElems('#LargeFirmware a.navlistsearch') startIdx = getStartIdx() # get firmware download URL for idx in range(startIdx, numFiles): assert files[idx].is_displayed() fileName = files[idx].text ulog('idx=%d, fileName="%s"'%(idx, fileName)) if 'firmware' not in fileName.lower(): continue prevTrail+=[idx] storeFile(modelName, files[idx]) prevTrail.pop() return PROC_OK except (StaleElementReferenceException): try: driver.find_element_by_css_selector("a.btn.close.fl-left").\ click() return TRY_AGAIN except (NoSuchElementException): return TRY_AGAIN except TimeoutException as ex: raise ex except Exception as ex: traceback.print_exc(); ipdb.set_trace() driver.save_screenshot('netgear_exc.png')
def main(): startCatIdx = int(sys.argv[1]) if len(sys.argv)>1 else 0 startFamIdx = int(sys.argv[2]) if len(sys.argv)>2 else 0 startPrdIdx = int(sys.argv[3]) if len(sys.argv)>3 else 0 global driver,conn harvest_utils.driver=getFirefox(dlDir) driver = harvest_utils.driver conn=sqlite3.connect('netgear.sqlite3') csr=conn.cursor() csr.execute("CREATE TABLE IF NOT EXISTS TFiles(" "brand TEXT," "category TEXT," "family TEXT," "product TEXT,"# -- is model "desc TEXT,"# -- is fileName "href TEXT," "file_sha1 TEXT," "PRIMARY KEY (product,desc)" ")"); conn.commit() driver.get('http://downloadcenter.netgear.com/') #click DrillDown waitClickable('#ctl00_ctl00_ctl00_mainContent_localizedContent_bodyCenter_BasicSearchPanel_btnAdvancedSearch').click() # # wait Page2 try: catSel=Select(waitClickable(catSelCss)) numCat=len(catSel.options) for catIdx in range(startCatIdx,numCat): catSel=Select(waitClickable(catSelCss)) print('catIdx=',catIdx) catTxt=catSel.options[catIdx].text uprint('catTxt='+catTxt) catSel.select_by_index(catIdx) waitTextChanged(famSelCss) famSel=Select(waitClickable(famSelCss)) numFam=len(famSel.options) for famIdx in range(startFamIdx,numFam): famSel=Select(waitClickable(famSelCss)) print('famIdx=',famIdx) startFamIdx=0 famTxt =famSel.options[famIdx].text uprint('famTxt='+famTxt) famSel.select_by_index(famIdx) waitTextChanged(prdSelCss) prdSel=Select(waitClickable(prdSelCss)) numPrd=len(prdSel.options) for prdIdx in range(startPrdIdx,numPrd): prdSel=Select(waitClickable(prdSelCss)) startPrdIdx=0 print("catIdx=%d, famIdx=%d, prdIdx=%d"%(catIdx,famIdx,prdIdx)) prdTxt=prdSel.options[prdIdx].text uprint('cat,fam,prd=("%s","%s","%s")'%(catTxt,famTxt,prdTxt)) prdWaiting = waitElem(prdWaitingCss) prdSel.select_by_index(prdIdx) WebDriverWait(driver, 5, poll_frequency=0.5).\ until(lambda x:prdWaiting.is_displayed()==True) WebDriverWait(driver, 60, poll_frequency=0.5).\ until(lambda x:prdWaiting.is_displayed()==False) #waitUntil(lambda:prdWaiting.is_displayed()==True) #waitUntil(lambda:prdWaiting.is_displayed()==False) numResults=waitText(numResultsCss,3) print('numResults=',numResults) if numResults is None: continue numResults=int(re.search(r"\d+", numResults).group(0)) if numResults >10: showMore=waitClickable("#lnkAllDownloadMore",3) showMore.click() try: erItems=getElems('a.register-product.navlistsearch',3) except TimeoutException: erItems=getElems('div#LargeFirmware > ul > li > div > p > a.navlistsearch',3) if len(erItems) != numResults: print('Error, numResults=%d, but len(erItems)=%d' %(numResults,len(erItems))) for erItem in erItems: if not erItem.is_displayed(): continue desc=getElemText(erItem) uprint('desc="%s"'%desc) href=erItem.get_attribute('data-durl') if not href: href=erItem.get_attribute('href') print('href=',href) if not href.startswith('http'): print('Error: href=',href) sql("INSERT OR REPLACE INTO TFiles" "(brand,category,family,product,desc,href)VALUES" "('Netgear',:catTxt,:famTxt,:prdTxt,:desc,:href)", locals()) uprint('INSERT ' '("%(catTxt)s","%(famTxt)s","%(prdTxt)s","%(desc)s","%(href)s")' %locals()) except Exception as ex: import ipdb; ipdb.set_trace() print(ex) import traceback; traceback.print_exc() print('-- terminate firefox') driver.quit()
def main1(catIdx, famIdx, prdIdx, executor): startCatIdx, startFamIdx, startPrdIdx = catIdx, famIdx, prdIdx driver = webdriver.PhantomJS() harvest_utils.driver = driver driver.get('http://downloadcenter.netgear.com/') # click DrillDown waitClickable('#ctl00_ctl00_ctl00_mainContent_localizedContent_bodyCenter' '_BasicSearchPanel_btnAdvancedSearch').click() # # wait Page2 try: catSel = Select(waitClickable(catSelCss)) numCat = len(catSel.options) for catIdx in range(startCatIdx, numCat): catSel = Select(waitClickable(catSelCss)) print('catIdx=', catIdx) startCatIdx = 0 catTxt = catSel.options[catIdx].text uprint('catTxt= ' + catTxt) oldText = getText(famSelCss) catSel.select_by_index(catIdx) waitTextChanged(famSelCss, oldText) famSel = Select(waitClickable(famSelCss)) numFam = len(famSel.options) for famIdx in range(startFamIdx, numFam): famSel = Select(waitClickable(famSelCss)) print('famIdx=', famIdx) startFamIdx = 0 famTxt = famSel.options[famIdx].text uprint('famTxt= ' + famTxt) oldText = getText(prdSelCss) famSel.select_by_index(famIdx) waitTextChanged(prdSelCss, oldText) prdSel = Select(waitClickable(prdSelCss)) numPrd = len(prdSel.options) for prdIdx in range(startPrdIdx, numPrd): prdSel = Select(waitClickable(prdSelCss)) startPrdIdx = 0 print("catIdx=%d, famIdx=%d, prdIdx=%d" % (catIdx, famIdx, prdIdx)) prdTxt = prdSel.options[prdIdx].text uprint('cat,fam,prd=("%s","%s","%s")' % (catTxt, famTxt, prdTxt)) prdWaiting = waitElem(prdWaitingCss) prdSel.select_by_index(prdIdx) try: WebDriverWait(driver, 1, 0.5).\ until(lambda x:prdWaiting.is_displayed() is True) except TimeoutException: pass try: WebDriverWait(driver, 5, 0.5).\ until(lambda x:prdWaiting.is_displayed() is False) except TimeoutException as ex: pass numResults = waitText(numResultsCss, 3, 0.5) if numResults is None: continue numResults = int(re.search(r"\d+", numResults).group(0)) print('numResults=', numResults) if numResults > 10: waitClickable("#lnkAllDownloadMore", 3).click() try: erItems = getElems('a.register-product.navlistsearch', 3, 0.5) except TimeoutException: erItems = getElems( 'div#LargeFirmware > ul > li > div > p > a.navlistsearch', 3) if len(erItems) != numResults: print('Error, numResults=%d, but len(erItems)=%d' % (numResults, len(erItems))) for itemIdx, erItem in enumerate(erItems): if not erItem.is_displayed(): print('itemIdx=%d is not displayed()' % itemIdx) continue desc = getElemText(erItem) uprint('desc="%s"' % desc) if 'firmware' not in desc.lower(): continue fw_url = erItem.get_attribute('data-durl') if not fw_url: fw_url = erItem.get_attribute('fw_url') print('fw_url=', fw_url) if not fw_url: continue if not fw_url.startswith('http'): print('Error: fw_url=', fw_url) continue executor.submit(download_file, prdTxt, desc, fw_url) # download_file(prdTxt, desc, fw_url) catIdx, famIdx, prdIdx = None, None, None return catIdx, famIdx, prdIdx except BaseException as ex: traceback.print_exc() dumpSnapshot('netgear_crawler.py.png') finally: driver.quit() return catIdx, famIdx, prdIdx
def fileWalker(): global driver,prevTrail try: waitUntil(isReadyState) prodName=waitText('#prodname') ulog('prodName="%s"'%prodName) modelName = waitText('#prodmodel') ulog('modelName="%s"'%modelName) # click "Support" retryA(lambda: elemWithText('li.tab-link', 'Support').click()) # expand "Downloads" waitClickable('div.accordion-section:nth-child(2) a').click() pageUrl=driver.current_url # select tables try: tables = getElems('table.supp',9,1) except TimeoutException: tables=None if not tables: ulog('no firmware download for "%s"'%modelName) trailStr=str(prevTrail) sql("INSERT OR REPLACE INTO TFiles (model,product_name," "page_url,tree_trail) VALUES" "(:modelName, :prodName," ":pageUrl,:trailStr)",glocals()) ulog('UPSERT "%(modelName)s", "%(prodName)s", '%glocals()) driver.back() return files = getElems('table.supp tr') startIdx = getStartIdx() numFiles=len(files) ulog('numFiles=%s'%numFiles) bUpserted=False for idx in range(startIdx, numFiles): try: col=files[idx].find_element_by_css_selector('td:nth-child(1)') except NoSuchElementException: ulog('bypass idx=%s'%idx) continue fwDate=guessDate(col.text) if not fwDate: ulog('bypass idx=%s'%idx) continue desc=files[idx].find_element_by_css_selector('td:nth-child(2)') fwDesc=desc.text fileName=desc.find_element_by_css_selector('a') ulog('fileName.text="%s"'%fileName.text) if 'firmware' not in fileName.text.lower(): ulog('bypass idx=%s'%idx) continue fwVer = guessVersion(fileName.text) fileUrl=fileName.get_attribute('href') fwDesc=desc.text.strip() trailStr=str(prevTrail+[idx]) ulog('trail=%s'%trailStr) sql("INSERT OR REPLACE INTO TFiles (model,product_name," "fw_date, fw_ver, fw_desc, " "page_url,file_url,tree_trail) VALUES" "(:modelName, :prodName," ":fwDate,:fwVer,:fwDesc," ":pageUrl,:fileUrl,:trailStr)",glocals()) ulog('UPSERT "%(modelName)s", "%(prodName)s", "%(fwDate)s", ' ' "%(fwVer)s", %(fileUrl)s'%glocals()) bUpserted=True if not bUpserted: trailStr=str(prevTrail) sql("INSERT OR REPLACE INTO TFiles (model,product_name," "page_url,tree_trail) VALUES" "(:modelName, :prodName," ":pageUrl,:trailStr)",glocals()) ulog('UPSERT "%(modelName)s", "%(prodName)s", '%glocals()) driver.back() return except Exception as ex: ipdb.set_trace() traceback.print_exc() driver.save_screenshot(getScriptName()+'_'+getFuncName()+'_exc.png')
def selectSoftwareType(): """ This page would be jumped to versionWalker() or either jumped back to treeWalker forward: may auto jump backward: not auto jump """ global startTrail,prevTrail,driver try: waitText('.csProductSelectorBreadcrumb', 5, 1) waitUntilStable('.csProductSelectorBreadcrumb', 1, 0.3) depth = getDepth() jumpedLevels =depth - len(prevTrail) ulog('jumpedLevels=%d'%jumpedLevels) assert jumpedLevels>=0 ulog('depth=%d, but prevTrail=%s'%(depth, prevTrail)) startIdxFromStartTrail=False def getStartIdx()->int: if startTrail: nonlocal startIdxFromStartTrail startIdxFromStartTrail=True return startTrail.pop(0) else: return 0 if jumpedLevels>0: while depth>len(prevTrail): startIdx=getStartIdx() prevTrail+=[startIdx] else: startIdx=getStartIdx() assert depth==len(prevTrail) ulog('startTrail=%s'%startTrail) ulog('prevTrail=%s'%prevTrail) ulog('startIdx=%d'%startIdx) ulog('url=%s'%driver.current_url) crumbs = waitText('.csProductSelectorBreadcrumb') uprint('crumbs=%s'%(crumbs.replace('\n',' > '))) if not hasElem('table#imageTableContainer', 1.5,0.4): if jumpedLevels>0: startIdx=getStartIdx() if depth > len(prevTrail): prevTrail+=[startIdx] sdpBannerTitle=waitText('td.SDPBannerTitle').strip() ulog('SDBBannerTitle="%s"'%sdpBannerTitle) assert sdpBannerTitle.lower().startswith('select ') waitUntil(lambda: getNumElem('div.csWrapper li a') > 0) swtypes = getElems('div.csWrapper li a') ulog('%s'%[(i,getElemText(_)) for i,_ in enumerate(swtypes)]) numSwTypes=len(swtypes) assert numSwTypes > 0 for idx in range(startIdx, numSwTypes): ulog('goto Trail=%s'%(prevTrail+[idx])) swtypes = getElems('div.csWrapper li a') ulog('Click "%s"'% getElemText(swtypes[idx])) clickElem(swtypes[idx]) prevTrail+=[idx] selectSoftwareType() prevTrail.pop() # Select a Product -> Select a Software type -> Select a Platform # https://software.cisco.com/download/type.html?mdfid=277873153&flowid=170&softwareid=283724313 # Downloads Home >Products >Cisco Interfaces and Modules >WAN Interface Cards >1700/2600/3600/3700 Series 2-Port Analog Modem WAN Interface Card >Analog Firmware Loader >Windows 2000-v6780 # not auto back to treeWalker # go back manually crumbs = getElems('.csProductSelectorBreadcrumb a') ulog('manually backto "%s"'%getElemText(crumbs[-1])) ulog('prevTail=%s'%prevTrail) clickElem(crumbs[-1]) # do I need to pop prevTrail? # prevTrail.pop() else: ulog('auto forward to versionWalker') if startIdxFromStartTrail: startTrail.insert(0, startIdx) for i in range(jumpedLevels): if not startTrail: break startTrail.pop(0) versionWalker() for i in range(jumpedLevels): crumbs = getElems('.csProductSelectorBreadcrumb a') ulog('manually backto "%s"'%getElemText(crumbs[-1])) ulog('prevTail=%s'%prevTrail) clickElem(crumbs[-1]) prevTrail.pop() except Exception as ex: ipdb.set_trace() print(ex); traceback.print_exc() driver.save_screenshot('cisco_selectSoftwareType.png')
def versionWalker(): global startTrail,prevTrail,driver try: waitClickable('.treeLinks > a:nth-child(1)') ulog('current_url=%s'%driver.current_url) crumbs=waitText('.csProductSelectorBreadcrumb').replace('\n', ' > ') ulog('crumbs=%s'%crumbs) # click Expand All numNodes = len(driver.find_elements_by_css_selector('.tree a')) ulog('number of versions=%d'%numNodes) try: with UntilTextChanged('.tree'): clickElem(waitClickable('.treeLinks > a:nth-child(1)')) except TimeoutException: pass treeText=waitText('.tree') ulog('treeText="%s"'%treeText) if startTrail: startIdx=startTrail.pop(0) else: startIdx=1 ulog('startTrail=%s'%startTrail) ulog('prevTrail=%s'%prevTrail) ulog('startIdx=%d'%startIdx) assert startIdx >= 1 try: prevFwVer=waitText('.tree a.nodeSel', 5) except TimeoutException: uprint("css='.tree a.nodeSel' not found") prevFwVer=None ulog('prevFwVer="%s"'%prevFwVer) for idx in range(startIdx, numNodes): nodes = driver.find_elements_by_css_selector('.tree a') isLeaf = (nodes[idx-1].text != '') ulog('goto Trail=%s'%(prevTrail+[idx])) if isLeaf: if not nodes[idx].text.strip(): continue fwVer=nodes[idx].text.strip() nodeClass=nodes[idx].get_attribute('class') ulog('fwVer="%s", nodeClass="%s"'%(fwVer,nodeClass)) if 'nodeSel' not in nodeClass: noWait= (prevFwVer==fwVer) if prevFwVer else False try: with UntilTextChanged('table#imageTableContainer',10,1,noWait): ulog('Click "%s"'%fwVer) clickElem(nodes[idx]) except TimeoutException: with UntilTextChanged('table#imageTableContainer',10,1,noWait): ulog('Click "%s" twice'%fwVer) clickElem(nodes[idx]) prevTrail+=[idx] tableRowWalker(fwVer) prevTrail.pop() prevFwVer=fwVer # go back page crumbs=getElems('.csProductSelectorBreadcrumb a') ulog('backto "%s"'%getElemText(crumbs[-1])) ulog('prevTail=%s'%prevTrail) if prevTrail==[2, 1, 0, 1, 2, 0]: ipdb.set_trace() clickElem(crumbs[-1]) except Exception as ex: ipdb.set_trace() print(ex); traceback.print_exc() driver.save_screenshot('cisco_versionWalker.png')
def tableRowWalker(fwVer:str): global startTrail,prevTrail,driver try: try: cells = getElems('table#imageTableContainer tr td') except TimeoutException: ulog("Timeout at getElems('table#imageTableContainer tr td'); bypass!") return assert len(cells)%4==0 numFiles=int(len(cells)/4) cellTxt = [getElemText(_) for _ in cells] try: carts = getElems('table#imageTableContainer tr td input', 10) cartsTxt = [getElemAttr(_,'title').lower().strip() for _ in carts] # https://software.cisco.com/download/release.html?mdfid=282822110&flowid=266&softwareid=280805680&release=15.0.2-SE8&relind=AVAILABLE&rellifecycle=MD&reltype=latest # fwVer='15.0.2-SE1(ED)' has len(cells)==6 and numFiles==4 cartsTxt = flatCarts(cartsTxt) assert len(cartsTxt)==numFiles needContracts=[ int('cart' in _) for _ in cartsTxt] # needContract=int(int(len(cells2)/numFiles)==2) except TimeoutException: ulog('needContract= "Deferral"') needContracts= [-1]*numFiles # Deferral ulog('needContracts=%s'%needContracts) quats = combinePerFour(cellTxt) assert len(quats)==len(needContracts) quats = list(zip(quats, needContracts)) fileDescs=[] for quat,needContract in quats: fileTitle=quat[0].split('\n')[0].strip() fileName=quat[0].split('\n')[1].strip() relDate=guessDate(quat[1]) fileSize=guessFileSize(quat[2]) fileDescs.append((fileTitle,fileName,relDate,fileSize, needContract)) model = waitText('td.SDPBannerTitle') ulog('model="%s"'%model) pageUrl=driver.current_url if startTrail: startIdx=startTrail.pop(0) else: startIdx=0 ulog('startIdx=%d'%startIdx) try: imageRhSide=driver.find_element_by_css_selector('div#imageRhSide') driver.execute_script( "arguments[0].scroll(0,%d);"%(62*startIdx), imageRhSide) except: pass spans=getElems('#imageTableContainer tr td span.overlay_img') assert len(spans)==numFiles for idx in range(startIdx, numFiles): ulog('getSha512 Trail=%s'%(prevTrail+[idx])) infos = getSha512(spans[idx]) if infos: infos.update(dict(needContract=needContracts[idx], model=model, pageUrl=pageUrl,treeTrail=str(prevTrail+[idx]))) else: fileTitle,fileName,relDate,fileSize,needContract=fileDescs[idx] infos=dict(fileTitle=fileTitle,fileName=fileName,relDate=relDate, fileSize=fileSize,needContract=needContract, pageUrl=pageUrl,treeTrail=str(prevTrail+[idx]), md5=None,sha512=None,model=model,fwVer=fwVer) sql("INSERT OR REPLACE INTO TFiles(model," " fw_date,fw_ver,file_title,file_name,file_size," " need_contract, page_url, tree_trail, md5, sha512) " "VALUES (:model," " :relDate,:fwVer,:fileTitle,:fileName,:fileSize," " :needContract, :pageUrl, :treeTrail, :md5, :sha512)", infos) ulog('UPSERT "%(model)s", "%(relDate)s", "%(fwVer)s", ' '"%(fileTitle)s", "%(fileName)s", %(needContract)d,' ' "%(treeTrail)s", %(pageUrl)s, "%(md5)s", "%(sha512)s"' %infos) except Exception as ex: ipdb.set_trace() print(ex); traceback.print_exc() driver.save_screenshot('cisco_tableRowWalker.png')